A Rapid Development Process for Marine

Propellers Through Design, Simulation and

Prototyping

by

© Xueyin Wu

A thesis submitted to the

School of Graduate Studies

in partial fulfi lment of the

requirements for the degree of

Master of Engineering

Faculty of Engineering and Applied Science

Memorial University of Newfoundland

October 20 10 Submitted

St. John's Newfoundland

Abstract
Flexibility and speed to market are the keys to successful product development. For

marine propellers, these goals are achieved though iteration of design software and

prototyping. In this thesis, an expanded propeller design code, "OpenPVL_SW",

which was developed based on the Open propeller vortex lattice lifting line code

(OpenPVL), was improved to include propeller geometry generation into SolidWork ,

thrust simulation with CosmosFioWorks and strength assessment with CosmosWorks.

The new code OpenPVL_SW is described in this thesis. In the OpenPVL_SW, a

parametric design technique and a single propeller geometry generator are completed

in MATLAB, and a propeller blade geometry file for SolidWorks is created after

running the design program.

The purpose of this study is to extend the use of the original OpenPVL code not only

for the propeller design, but also to achieve the thrust simulation and strength check

using SolidWorks tool package, CosmosFloWorks and CosmosWorks. A propeller

geometry, which is provided by Oceanic Consulting Corporation, is simulated to

predict propeller thrust using CosmosFloWorks. A case study designed an AUV

propeller and simulated the thrust with CosmosFloWorks to prove that the

OpenPVL_ SW code provides perfect propeller geometry and a reasonable simulation

result of thrust.

11
Acknowledgements
First and foremost, appreciation and thanks are bestowed upon Prof. Andy Fisher for

his guidance, support and advice. From the beginning stages to the submission of this

thesis, he supported and kept me on the path towards completion, and I did learn a lot

from him. I am greatly indebted.

I expresses thanks to Matthew Garvin, manager of design and fab1ication in Oceanic

Consulting Corporation, for providing propeller geometry, experimental results and

guidance.

I would like to extend thanks to Prof. Mahmoud Haddara. I learned a lot from him

about propeller design and testing.

I would like to acknowledge the support of the people in the rapid proto typing lab for

their support during the propeller fabrication using Stratasys Fusion Deposition

Modeling (FDM).

lll
Table of Contents

Abstract ... ....... .. ..... ............. ... ......... ...... ....... .............. ..... ......... ... ... ........ .. ....... .. .. .. .. .. ..... .... ... ... ... ii

Acknowledgements .. .... ..... ..... ... .. ... .... ........... .... ... ....... .. .... ....... ..... ......... ..... ..................... ...... .. . iii

Table of Contents ......... ........... .. .. .................. .... .......... .. ............ ..... ... ..... ..... ...... ... ........ ..... ........ iv

List of Tables ... .. ..... ....... ........... ..... .. .. ............ ..... ........ .... ....... .. ......... .. ............. ......... ...... .. .. ... ... vi

List of Figures .... ... ... .................. ... ................. .. .. ............. ...... .. .. .. .. ........ ... .... .... .. .. .... ... ...... .. ... .. vii

Introduction .. ..... ..... .. ....... ..... .... .... ................ .. ...... .............. .. ... ...... ... .. ... ....... .. .. .. ..... ..... .... ... . 1

2 Literature and Software Review ........ .............. .......... ..... .. ... .. ..... ..... .. ... .. ..... ........... oo ..... oooo···· 5

• 2.1 Review of Propeller Design Methods .. ........ 00 . . . . . . . . . . . . . . 00 00 . . 00 00 . . . . . . 00 . . 00 . . . . . . . . . . . . . . . . . . 00 00 5

• 2.2 Review of Recent Propeller Design Software .......... 00 . . . . . . 00 00 00 . . . . . . 00 . . . . . . . . . . . . . . . 00 00 . . 0 0 . 10

• 2.3 Review of OpenPYL 00 . . . . . . . . . . . . . . . . . . . . . . ... . ......... .... ........ ...... .. 00 . . . . . . . . . . . . . . . . . . . ............... 13

• 2.3 .1 Parametric Analysis .. ... ........ ..... .. .. ............ 00 . . . . . . . . . . . . . . . . . . . . . . . ... . ..... . .... . . ........... .. 13

• 2.3 .2 Blade Design ... ......................... ...... ... ............ ... .. .. ........ .................... ......... .... . 19

• 2.3.3 Propeller Geometry Development Through CAD .. ...... .. ................. .......... .. .... 26

• 2.4 Review of Computer-Aided Design (CAD) .......... ............ ........... .. .. .. .................... 28

• 2.5 Review of Computational Fluid Dynamics (CFD) and Finite Element

Analysis (FEA) .... ... ... ... ... .......... ............ ....... .. .. ... .. ........ ............... ... ............ ... .... 30

• 2.6 Review of Propeller Fabrication .......... .. ....... ........... .. .. .... .. .. .. ...... ........ ... ............ ... 32

• 2.6.1 Traditional Fabrication Processes ........ .. ...... .. .......................................... .... ... 32

• 2.6.2 Rapid Fabrication Processes .... ...... .. 00 00 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . oo .. 0000 . . . . 00 .... 00 34

• 2.7 Review of Propeller Testing .00 . . . . . . . . . . ...... . . .................. . . ..... ......... ..... .. .. . . . ........ . ... . .. 38

3 Marine Propeller Design and Simulation in SolidWorks .... .. ..................... 00 . . . . . . . . . . . . . . . . . . . . . . 43

• 3.1 Open_PYL and OpenPYL_SW Codes ...... .................................. ............ .. .. .... .. ..... 44

IV
 • 3.2 Using Solidworks for Propeller Blade Geometry ................................................... 47

• 3.2.1 Propeller Foil Sections ........................ ........ .................. .. ..... .............. .... ... 48

• 3.2.2 Propeller Blade Surface Development. .. ................................ ..... .. ..... ........ 56

• 3.3 Simulations using CosmosFloWorks ..... .... ......... .. ............. .. .............. .... ........ ..... ... 60

• 3.4 Simulations using CosmosWorks ......... ............... ............... ... .......... .............. .... .... 66

4 Case Study: Simulation of an AUV Propeller. ..... .... ........ ........ ........ .... ..... .. ......................... 76

5 Case Study: Simulation of a Marine Propeller Provided by Oceanic Consulting

Corporation ... .. .. ....... .. ........ ....... ........ .......... ...... ..... ......... .......... ..... ........... .. ... 88

6 Propeller Fabrication by Rapid Proto typing ....................... ..... .............. .......... .................. .. 98

7 Conclusions and Future Work ..................... ....... ...... ........ .. .. ........... ..... .... ............... .......... 105

• 7.1 Conclusions ......... ..... .... ... .. :............................. ..... ... ...... ... ....... ........... ...... .. ..... .... I 05

• 7.2 Future Work ............. ... ............. ... .... ......... ... ...... ............ ......... .. .............. ... .. ..... .. 106

References .. ...... ...................... ............... ...... ....... ..... .... ... ... .......... ..... ....... ............................... 107

Appendix A: The Part ofOpenPVL_SW MATLAB Source Code fo r Generating

Propeller Geometry in SolidWorks ... .... ..... .......... .... ..... ..... .... ..... .... ... ......... .. .. .... 111

Appendix B: SolidWorks Macro for the AUV Propeller Geometry: ......... ............ ........... ........ 11 9

v
List of Tables
Table 1: Required Input Data for Lifting Line Method and Panel Method .. .... ... .... ........ ..... .... ... .. 9

Table 2: Advantages and Disadvantages of Each NACA Series ... .... ... ...... .. ... ... ......... ............. .. 22

Table 3: Prototyping Technology Materials .... ............... .......................... .. .. ....... ... ...... ....... .... .. 35

Table 4: Processes of Propeller Design by Open_PVL and OpenPVL_ SW ..... .. .. ....... .. ............. 46

Table 5: The information ofNACA 63A006 and mean line a=0.8 (modified) ....... .... .......... .... .. 55

Table 6: Simulation Results of the Three Propellers ..... .......... ..... .. .. .... ....... .. ... .... .. ........ .... .. .... .. 83

Table 7: StrataSys FDM 2000 Specifications ....... ...... ......... .. ... .. ... ... .. .... .... ... .. ...... ... ................. 99

VI
List of Figures
Figure 1: Flow Chart of Propeller Design by Computer Programs ................................. .... .......... 3

Figure 2: Parametric Analysis Matlab Interface .. .. ............................ ........ ..... ......... ....... ........... 14

Figure 3: Efficiency Plot. .... .............. ... .... .... ... ... ........ ....... .. ... ........ .................... .. ... ... ........ ..... .. 19

Figure 4: Propeller Design Matlab Interface ... .... ..... .. .......... ....... .......... .......... ........ .. .............. .. 20

Figure 5: Foil Section Geometry ..... ..... ..... ... ............. ...... ........... .. .................................. ........... 23

Figure 6: Graphical Report ···· ·· ········· ········· ··········· ········ ···· ················ ·· ·· ·· ··· ················:· ··· ·········· 24

Figure 7: 2D Propeller Blade Profile ... ..... .... ....... .... .... ........ .. ..... ....... .............. .. ....... ....... ....... ... 25

Figure 8: 3D Propeller Image ... ......... .. ........ ........ .... ... ......... ......... ......... ........................ ... ..... .... 25

Figure 9: Propeller in Rhino ................ .... ........ ..... ...................... ........................... .......... .... ..... 27

Figure 10: Propeller Geometry Validation ........ .. ............... ........ ......................... ... ...... .... ... ... ... 28

Figure 11 : Flow Chart of Casting Processes .... .. ........... ..... .... ........... ... .. ......... ..... ..... ......... ..... .. . 33

Figure 12: Wind Tunnel Model Constructed by FDM ................ .. ... ....................................... ... 36

Figure 13: Sketch of a Propeller Testing Set-up ........................ ..... .. .. ...... ....... .. .... ...... ..... .. ..... .. 38

Figure 14: Thruster Mount. .... ........ ..................... ....... ..... .... ............ ........... .... ... ... ..... .. ...... ... ..... 39

Figure 15: Tow-tank Carriage .................. .. ........ .... ..... .... ............. ................ .. .......... ........... .. .. .. 40

Figure 16: Flow Chart of Generating a Propeller Blade ......... ... ..... ....... ... ...... .... ..... .... .. ......... .... 48

Figure 17: Foil Section Points .... .. ....... ............... ..... .......... ........ ..... ..... ........ ... ... ..... ............. .. .... 49

Figure 18: A Propeller Foil Section ... ...... .. ..... .... ....... ....... ......... .. ....... .. ..... .. ..... ...... .................. 50

Figure 19: Overview of Propeller Foil Sections ................................ .................... .............. ...... 51

Figure 20: Sideview of Propeller Foil Sections ... ..... ... ...... .. .. .... ......... .. ... .... ......... ......... ..... .. ..... 52

Figure 21: Foil Section using NACA 65A01 0 a=0.8 at the r/R=0.35 ................... ...................... 54

Figure 22: Foil Section Using NACA 63A006 a=0.8(modified) at the r/R=0.35 ..... ....... ...... ... .. . 56

Figure 23: Propeller Foil Sections and the Leading Edge Line ... .. ..... .. ..... .. ..... .... ......... ..... ........ 57

Vll
Figure 24: A Propeller Blade ... .... ....... ... .. ... ................... ............ .. ........ ...... .. .. .. .... ... .... .... .......... 58

Figure 25: A Propeller Blade with a Hub and a Nose Cone ........ .. .............. .. .................. .. ...... ... 59

Figure 26: A Propeller with Three Blades ........ .............. ... ....... .. ... ..... ... ........ .... ........... ..... ........ 60

Figure 27: Simulation Domain of Propeller .............. .. .... .. ............... .. .................................. .. ... 62

Figure 28: Pressure Results of a Propeller Simulation with Domain Size 6mx6mx30m

(widthxdepthxlength) ............. ... .... ................................ .......... ......... ........ ...... ... 63

Figure 29: Pressure Results of a Propeller Simulation with Domain Size 4mx4mx20m

(widthxdepthxlength) .................... ... ................................ ...... .... .. ......... ..... ... .... 64

Figure 30: Side View of a prope ller ................................ .... .................. .. ...................... ....... .. ... 65

Figure 31: A Simulation Model Without Blades ............................ .. ........ .. ...... .................... .... . 66

Figure 32: Surface Total Pressure Distribution of a 3-Blade Propeller...... .. ........ ...... .. ........... .. .. 69

Figure 33: Dynamic Pressure Distribution of a 3-Blade Propeller .......... ................ .... .......... ..... 70

Figure 34: Stress Distribution of the A UV Propeller ........ ..................................................... .... 71

Figure 35: FOS of the AUV Propeller. .......... .... .... ... .. ...... .... ........ .... ............ ................ .. ....... .... 72

Figure 36: Fillets at the Root of a Propeller Blade .................... .. .. .............. ............................... 73

Figure 3 7: Stress Distribution with Fillets of the 3-Blade propeller ...... .... .. .. .......... ................... 73

Figure 38: FOS distribution with Fillets of the 3-Blade Propeller .. .. .. .... .......... .. ............ .. .. ...... .. 74

Figure 39: Flow Chart of a Propeller Simulation Process .......... .... ........ ........ ........ ...... .............. 75

Figure 40: AUV Propeller Blade Geometry ............ .......... .............................................. .. .... .. .. 77

Figure 41: Final AUV Propeller .......... ..... ... ............. ... .............. .... ......... ......... ...... .......... ........ .. 78

Figure 42: Predicted Result of Propeller Thrust With Blades .............. .... ............ .............. ...... .. 79

Figure 43: Predicted Result of Propeller Thrust Without Blade .. .... .... ....................................... 80

Figure 44: Relation of RPM and thrust at vessel speed 1rn/s .... ...................................... ........... 81

Figure 45: Plot of RPM and Thrust from Dr. Vural's Experiments .................. .......... .. .... ...... .... 81

Vlll
Figure 46: Surface Dynamic Pressure Distribution of the AUV Propeller .... .......... ..·.... ........ ..... 84

Figure 47: Stress Distribution of the AUV Propeller. ... ..... .... ...... ... ...... ......... .... .... .... .. ..... ....... .. 85

Figure 48: Deflection of the AUV Propeller. ... .. ............. ... ... ........ ..... .... .. ... .. .. ... .... ........... ........ . 86

Figure 49: FOS of the AUV Propeller .... .... .......... .... .... .. ........................ .. ...... .... ........ .. ........ ..... 87

Figure 50: Propeller Geometry-4 Blades .... .. .. ... .. ........ ................. .. .. ... ........... .... ............ .... ..... .. 89

Figure 51 : Sideview of the Propeller-4 Blades .... ..... ......... .... ... .......... .. ............ .... ............. ...... .. 89

Figure 52: Predicted Propeller Thrust - 4 Blades .. ...... ..... ............ .. ......... .. .... .......... .... .. ........ .. .. 9 1

Figure 53: Propeller Geometry without blade .......... .. ................ .. .......... ........ .......... .... .... .......... 92

Figure 54: Predicted Propeller Thrust for the Propeller without Blade ...... .. .... .. .... .............. ...... . 93

Figure 55: Surface Dynamic Pressure of the Propeller-4 Blades .... .... .. .... ........ .. .. ...... .... .. .. .... .... 94

Figure 56: Stress Distribution of the Propeller-4 Blades .. .. .. ...... .... .... .. .. ........ .. .. ............ .......... .. 94

Figure 57: Deflection of the Propeller-4 Blades .. .... .... ........ .. .................................................... 95

Figure 58: FOS of the Propeller-4 Blades .. ............. ...... .... .... ........................ .... .... ....... ...... .... .. . 96

Figure 59: Flow Chart of the Procedures of a Propeller Design and Simulation using

OpenPVL_ SW, SolidWorks, CosmosFloWorks and CosmosWorks .................. .. .... 97

Figure 60: Stratasys FDM 2000 ..... ... .... .... ........ ...... ... .... ... .. ....... ... ..... ... ..... ..... ... .. .. ... ..... ....... .. 100

Figure 61: Screenshot of Propeller Blade Generated by Stratasys Quicks lice Software ... .... .... 101

Figure 62: A Propeller Fabricated by FDM .... .................... .. .... .......... ........ ........................ ..... 103

IX
Chapter 1

Introduction
In Newfoundland and Labrador, ocean technology is an important sector with many

public and private firms involved in designing new or more efficient technologies. One of

the main components of many of these developments is the marine propeller. Designing

high efficiency and low cost propellers for ships, AUV's, ROV's and other marine

applications is therefore a critical research topic. At the same time, the technologies of

Computer-Aided Design (CAD) and Computer-Aided Manufacturing (CAM) have been

rapidly advancing, making these tools much more accessible. More and more designs and

fabrication processes are using computer technologies to achieve high accuracy and

shortened cycle times for manufacture and testing. In this thesis, the focus is on a rapid

development process for marine propeller design and prototype fabrications using these

computer technologies.

Propellers transmit power by converting rotational motion into thrust. Traditionally,

engineers needed to do a lot of analysis to detennine the foil sections, pitch, blade angle

and so on, in order to design and produce a suitable propeller. The process took engineers

1
a long time to analyze, draw, fabricate and test. owadays, due to rapid advances in the

development of computer technologies, most engineers are using computer programs to

assist in the design of propellers, instead of using traditional methods. After designing a

propeller in a computer, rapid prototyping technologies provide a very efficient way to

fabricate a prototype propeller. The material that can be used in the rapid prototyping

technologies is not perfect for the actual propeller, however, using rapid prototyping

machines, prototype propellers can be produced quickly facilitating early testing. If the

results show the propeller is not perfonning as desired, the propeller can be redesigned,

reproduced and retested. Saving time is the main advantage of the rapid proto typing

technologies.

The general processes of propeller design and fabrication is to first input the propeller

parameters, generate a preliminary geometry and then to run a computer program to do

analysis. If the analysis results show that the propeller is suitable then the propeller

geometry outputs will be used for propeller fabrication and testing; if results show the

propeller is not desired, engineers will change propeller parameters to redesign it. Figure

1 shows a flow chart of the proposed method of design and propeller fabrication that will

be described in this thesis.

2
 Start with
Produce
vessel
propeller
mission
profile
Ye
No
Extract
propeller Acceptable
design test data
parameters

Calculate propeller Tow-tank

blade geometry Testing; model
based on propeller testing method
design method (large size),
direct testing
(small size)
Import propeller
blade geometry in
CAD software Generate STL
file and rapid
prototype

Design a hub and Ye

complete propeller No
geometry; choose Acceptable
propeller material strength

Use Computational
No Use Finite Element
Fluid Dynamics
Analysis (FEA) to
(CFD) to predict
check strength.
propeller thrust
Compare the
predicted and
desired thrust
Yes

Figure 1: Flow Chart of Propeller Development

3
Simulation work is now commonly done before a propeller is fabricated . If an integrated

software tool can include analysis, design, fabrication geometry and simulation, the

design process will be much more convenient for developers.

In this thesis, the author is expanding an open source propeller design code to

OpenPVL_SW code that is combined with SolidWork software, to analyze, design and

fabricate propellers, and also to see how the propeller design can be simulated. A case

study of an autonomous underwater vehicle (AUV) propeller completed by the

OpenPVL_SW code will be presented in this thesis. In this case, the OpenPVL_SW code

is used to create an AUV propeller and generate data for propeller geometry and

simulation. Second case study used CosmosFloworks to simulate the thrust of the

propeller where the geometry was provided by Oceanic Consulting Corporation.

Compared with the real testing result of the thrust, this case study is used to prove that

CosmosFloworks can provide a reasonable thrust prediction. After the thrust simulation,

CosmosWorks is used to check the strength of the propeller deisgn using Finite Element

Analysis (FEA). If the strength result is suitable, then the propeller geometry is ready to

be fabricated; if the strength result is not desired, engineers can use higher strength

material or redesign the propeller geometry to increase strength.

4
Chapter 2

Literature and Software Review

This chapter will review the literature concerning propeller design methods, recent

propeller design codes, Computer-Aided Design (CAD) and Computer-Aided

Manufacturing (CAM) for propellers and also provide a review of OpenPVL.

2.1 Review of Propeller Design Methods

The development of propeller design methods has evolved for a long time. When

propellers were first used for propulsion, very few people knew how propellers operated

and the optimal way to design them. Early designs were following the steps: trial, error

and imagination. Momentum theory was applied to propellers in the late nineteenth

century. It explained the resulting thrust of propellers, but it did not give a detailed reason

for these results [1]. Design for blade strength was based on experience and then later on

simple beam theory to determine the minimum root thickness and thickness distribution

along the blade span. Nowadays, finite element analysis is commonly used to analyze

propeller blades in details to ensure structural soundness.

5
Early theoretical analysis applied to propellers was based on momentum theory. In 191 0,

Betz was the first to formally formulate the circulation theory of aircraft wings for using

with screw propellers [2]. The Vortex theory of propellers was developed with the basic

assumption that certain geometric qualities of the flow in which a propeller operates must

exist ifthe energy losses are to be minimized. In 1929, Goldstein developed this theory,

showing that the flow past a vortex sheet could be calculated by relating the two

theoretical cases of a propeller with a finite number of blades and a propeller with an

infinite number ofblades [3]. Goldstein formulated Betz's vortex theory for propeller for

real cases of propellers with a finite number of blades. He showed that the velocity

characteristics changed substantially by removing the assumption of the infinite blade

number. His analysis was done for the case of optimum circulation distribution of a

propeller with minimum energy loss. By calculating the ratio of the circulation with

infinite and finite number of blades, the Goldstein coefficient, K, was derived for periodic

flow [3].

In 1947, Glauert published that the individual airfoil sections along the blade span could

be included directly into a lifting line calculation to design a propeller owing to certain

characteristics of lift and drag. He used momentum theory to determine the average

induced velocity at the lifting line [4].

In 1952, Lerbs simplified the calculations by introducing the concept of induction factors

[5]. The induction factors depend only on helix geometry and therefore can be calculated

6
dependent of loading. To achieve this, Lerbs made two assumptions. One, that the radial

induced velocity is assumed to be negligible in the cases of light to moderate loading; the

other, that the calculated hydrodynamic pitch approximates the shape of the streamlines

in the wake. These are assumed to depend only on the axial and tangential velocity

components [5]. Lerbs extended Glauert's methods to calculate element contributions to

thrust and torque along the blade span. The results were integrated over the propeller

blades to give overall performance of the propellers. He effectively produced the first

·practical numerical propeller design method applied under various non-optimum

conditions of loading [5]. Lerbs' method became known as the lifting line method for

marine propellers, because the force produced by the circulation is assumed to act along

radial lines in place of the blades, similar to the theory of airscrews [6].

In 1955, Eckhardt and Morgan developed an engineering approach to the lifting line

method. They assumed an optimum distribution of circulation, and a reduced thrust,

proportional to the Goldstein factor due to the number of blades, and proceeded to

calculate induced velocities. This greatly reduced the amount of calculations. They also

showed that their method produced only small errors under light and moderate loading

conditions and therefore it is a very useful design tool [7].

In 1976, panel method was initially developed as lower-order method for incompressible

and subsonic flows [8]. In 1985, Hess and Valarezo introduced the panel method for

marine propellers that allowed them to vary the pitch values and distributions and take

into account the inflow wake distribution and cavitation effects [9]. The panel method

7
provides an elegant methodology to solve a class of flow past arbitrarily shaped bodies in

both two and three dimensions. The basic idea is to discretize the body in terms of a

singularity distribution on the body surface, to satisfy the necessary boundary condition ,

and to find the resulting distribution of singularity on the surface, thus obtain fluid

dynamic properties of the flow [ 10].

Nowadays, engineers are seeking ever faster ways to design propellers for customers'

requirements and enhancing speed to market. The lifting line method and panel method

are very useful and commonly used to design conventional propellers. The required input

data for lifting line method and panel method is shown in table 1.

8
 Table 1: Required Input Data for Lifting Line Method and Panel Method [ 11] [ 12]

Input data for lifting line method Input data for panel method

Number ofblades Number of blades

Propeller speed Propeller speed

Propeller diamete Propeller diameter

equivalent
Required thrust Required thrust

Ship velocity Ship velocity

Fluid density Fluid density

Maximum iterations in wake alignment Effective power

Hub vortex radius/hub radius Ship advance coefficient

Shaft centerline depth Wake fraction

Inflow variation Thrust deduction fraction

Ideal angle of attack Propeller advance coefficient

Hub diameter Required resistance

Propeller advance speed

Required thrust coefficient

Required torque

Required torque coefficient

Required propeller efficiency

Delivered power (KW)

9
As can be seen in Table 1, the lifting line method does not require as much detail as the

panel method, thus it can be obtained more easily and can still yield useful results [ 11].

By using the lifting line method, large advances in the design of marine propellers can be

made without much more detailed numerical input required. The lifting line method is

very useful for conventional propeller design and is used extensively by leading propeller

manufacturers [6]. By using this method, the circulation distribution and hydrodynamic

pitch can be calculated, and the required pitch distribution can be constant or varied.

Blade cross-sectional characteristics are incorporated into the calculations by using

detailed airfoil section data and lift-drag relationships. A strength analysis can also be

included in the method by incorporating simple beam theory directly into the calculation

to obtain the primary stresses [13] [14], which include bending moments due to thrust and

torque. Equation 1 is the bending moment at the section due to the thrust on the blade.

Equation 2 is the bending moment at the section due to the torque on the blade [ 15].

R 1 dT
Mr =
i"Z- dr-(r -r )dr
0

R 1 dQ
M =
Q I -rZ- -dr( r - r )dr
'h
0 2

Tis thrust; Q is torque; R is the propeller radius; r0 is hub radius; Z is the number of

blades.

2.2 Review of Recent Propeller Design Software

Based on many years of development, propeller design methods have been enhanced. Due

to the rapid development of computer technology, propeller designers are tending to u e

10
computer code more often to design propellers, and are designing much more

complicated and multifunctional codes. In the early stages of propeller design software,

the code was used for specific functions. For example, the WAOPTPROP code [16] was

only used for propeller geometry, and the PPT2 code [17] was only used for propeller

analysis. Later, computer code, such as the PVL code, was developed to design the

propeller's geometry and simultaneously analyze the performance [17]. Subsequently,

engineers wanted to combine computer code with CAD software to automatically fini h

the propeller's fabrication. For example, D'Epagnier created an OpenPVL code [11] not

only for analysis and design, but also to create scripting for 3D printable files using the

CAD software RHINO, which generates a .STL file that is used to build propellers by

rapid prototyping machines.

In 1991 , Hofinann wrote the WAOPTPROP code using VAX FORTRAN based on lifting

line method [ 16]. The program calculated the induced velocities at the blade sections

from a non-optimum circulation distribution, the required pitch distribution, and the thru t

and torque coefficients for the design condition. Propeller design used these program

results to determine the hydrodynamic pitch distribution by comparing the design point

coefficient with the pitch distribution. W AOPTPROP can also provide infonnation about

final propeller geometry. PPT2 was another program written by Hofinann [ 16] for the

analysis of propeller performance. Finally, W AOPTPROP combined with PPT2 can work

as a fully integrated propeller design program. However, this design program has two

main disadvantages: it needs two separate programs to achieve the design target, whereas,

it could be more convenient to use one program. Another disadvantage is that the final

11
 - - - - - - -- - -- - - - - -- - -- ----------

propeller geometry has to be entered by hand into CAD software, which is troublesome

when the designer has many priorities.

PVL code, was created by Professor J. Kerwin at MIT in 2001 using Fortran

programming [ 17]. It was translated into MATLAB as an open source MPVL code

released by Hsin-Lung in May 2007. MPVL code was based on lifting line method and

its optimization algorithm was based on Lerb's criteria [ 18]. MPVL code, working with

the high-level technical computing language MATLAB, can be easily modified by users

according to their specific needs; propeller designers are able to conduct both propeller

analysis and single propeller design functions of the program. Compared with Hofmann's

propeller design programs, MPVL code has the advantages of integrating design and

analysis programming. However, there is no way to connect the propeller design program

directly with CAD software, which means that even by using this code, users have to

create the propeller geometry in the CAD software by hand.

In September 2007, D'Epagnier [11] modified the MPVL code to create a new program

called OpenPVL. It operated using an evolved MPVL code while expanding upon

MPVL's applications, and had been modified to create scripting for 3D printable files

using a CAD interface to the commercial CAD software "RHINO". OpenPVL made up

for the disadvantages of MPVL by connecting the program to CAD software. In

OpenPVL, propeller blade geometry can be generated and imported into RHINO, and

then saved. Once in RHINO, various CAD manipulations are possible, including the

12
export of stereolithography (.STL) file, which can be used in a rapid prototyping (RP)

machine to produce propeller prototypes.

2.3 Review of OpenPVL

OpenPVL is an open source for marine propeller design. There are two main components

of OpenPVL. One is parametric analysis, which is used to combine all the propeller' s

parameters to analyze the efficiency and optimize the design. The other component is the

propeller design function to generate files of propeller inputs, outputs, geometry and

performance. The CAD software RHINO opens the geometry file and can from this one

use standard CAD commands to create a propeller blade, design a hub and add other

blades to complete the propeller.

2.3.1 Parametric Analysis

The number of blades, the propeller speed and the propeller diameter work as the three

foundational parameters for propeller design. These three are combined with other

additional parameters, including required thrust, ship speed, hub diameter, number of

vortex panels over the radius, maximum number of iterations in wake alignment, ratio of

hub vortex radius to hub radius, hub and tip unloading factor, swirl cancellation factor,

water density and hub image flag, as input parameters to the analysis process.

13
 Min wax Increment P Hub Image Flag {Check for YES)
Number of Bled. .
Propeller Speed (RPM)

Propeller Diameter (m)

I ..0000 Requ~red Thrust (N)

~ Sh•pVwlociiJ(m/1)
r-o.r- Hub Diameter (m)
~ Number ofVortu Panels overt he Radius
~ Mn. Heraliont in Wah Alignment
~ Hub Vortex Rdius/Hub Rtdiue

~ Hub Unlading Factor: D=a:Optimum

~ npUnloading Faclotr. 1• Reduced Loading
~ Swirl Canullalion F.ctor. 1z:No Cancellation

~ Water Oenaity (kllfm-'3)

Run OpenProp I

Figure 2: Parametric Analysis Matlab Interface

These input parameters are introduced in the sections below. Figure 2 shows the

parametric analysis, which includes the user input fields required to run the analysis.

• Number ofblades: The range of the number ofblades is from two to six.

Propellers normally have a number of blades within this range. More number of

blades will increase thrust; however, it may cause cavitaion. Less number of

blades will avoid cavitaion; however, it decreases thrust.

• Propeller speed: The unit of propeller speed is revolutions per minute (RPM). The

restrictions are that the value must be positive, and the maximum value must be

greater than or equal to the minimum value. The increment cannot be negative.

High RPM will increase thrust; however it may cause cavitation. Low RPM will

14
 - ------------------------------------------------------

avoid cavitation; however, it decreases thrust. The range of RPM cannot exceed

the rotational range by vessel engine.

• Propeller diameter: The unit of propeller diameter is meters (m). The value of

propeller diameter must be positive, and the maximum value must be greater than

or equal to the minimum value. The increment cannot be negative. The propeller

diameter must be greater than the hub diameter. Large propeller diameter

increases thrust, and small diameter will decrease thrust. The propeller diameter is

limited by the geometry of vessels.

• Required thrust: The unit of required thrust is Newtons (N). The value must be

greater than 0, and cannot be negative. The required thrust is deri ved based on the

required ship speed and resistant force when a vessel is moving.

• Ship speed: The unit of ship speed is meter per second (m/s). The value must be

greater than 0, and cannot be negative.

• Hub diameter: The unit of hub diameter is meters (m). The hub diameter has to be

greater than 15% of the propeller diameter [ 18].

• Number of vortex panel over the radius: This input field represents how many

vortex panels will be divided into the blade and thus affects the resolution of the

15
 - - - · - - - - - - - -- - - -- - - - - - -- - - -- - -

propeller blade. The number is greater than 0, and is an integer. Usually, twenty

panels can provide sufficient resolution [ 18].

• Maximum number of iterations in wake alignment: This input field determines

how many iterations Open_PVL is allowed to align the wake. The number is

greater than 0, and is an integer. Ten iterations are usually sufficient for the

program to converge and align in the wake [ 18].

• Ratio of the hub vortex radius to the hub radius: The hub drag was computed as a

function of the ratio of vortex core radius to hub radius, but the precise value of

this ratio is not critical [17]. For convenience, this ratio was usually assumed to be

one[18].

• Hub and tip unloading factors: These two factors are defined as the fractional

amount that the difference between the optimum values of tan Pi and tan Pare
reduced. If hub unloading factor is 0, tan Pi- tan p at the hub is retained at its
optimum value from Betz/Lerbs criterion. If hub unloading factor is I , tan Pi - tan
p at the hub is set to zero, and the values up to the mid span of the blade are
blended parabolically to the optimum value. The same procedure applies to the tip

[ 17].

16
 -- - - - - - - - - - -- -- -- - -- - - - -- - - - - - - - -

• Swirl cancellation factor: The value of this factor is zero or one. The swirl

cancellation factor is zero for contra-rotation propellers in which the tangential

velocities from each blade were cancelled [17). If there was no swirl cancellation,

the value of this factor was one [ 17].

• Water density: The water density depends on the users' preference. The unit is

kg/m 3 . The default value is 1025 kg/m 3 .

• Hub image flag: this input field is located on the upper right side. By checking this

option, hub image is present, and the circulation has a finite value at the hub. Un-

checking this option will have a zero circulation at the root of a propeller blade

[ 18].

On the right side of the parametric analysis screen, there are five editable parameters. r/R

is the ratio of the radial location to the total length of the blade radius to be set from 20%

of propeller radius (r/R=0.2) to the blade tip (r/R= l). c/D is the non-dimensional chord

distribution over the radius. Cd is the drag coefficient over the radius. Vs is to the

advance velocity over the radius. Va/Vs is the ratio of the axial inflow velocity to the

advance velocity over the radius. The value of VaNs is set to one to assume uniform

inflow. VtNs is the ratio of the tangential inflow velocity to the advance velocity over the

radius. The value ofVt/Vs is set to zero to assume uniform inflow. Open_ PVL

automatically interpolates these input fields in accordance with the hub radius.

17
. - - - - - - - - - - - - - - - - - - - - - - - - - - - ---

After the calculation process is completed, the program also creates the efficiency curves

according to the blade number and propeller diameter, shown in Figure 3. Ideally, a good

propeller has a large diameter, slow speed, low number of blades and high efficiency.

However, the real propeller parameters are always restricted in size and speed. It is the

purpose of the efficiency curves combined with different propeller diameters and speed to

help designers to determine the optimum parameters for a propeller design. To clearly

show the design approach for a propeller, Figure 3 is taken as an example. Due to the

limitation of a ship body geometry, the propeller diameter is restricted to three meters.

The ideal number of blades is small; however, less number of blades perhaps causes

propeller vibration due to the increased thrust on each blade. For this reason, a four-blade

design is adopted for this design. From the efficiency curves with three-meter diameter

and four-blade, the propeller has the highest efficiency at 100 RPM. Finally, the propeller

is chosen with three-meter diameter, four blades and 100 RPM. Now, the three major

propeller parameters have been determined. The detailed blade design is ready to be

conducted next.

18
 Number of Blades: 3 Number of Blades: 4

g 0.4
Ql

~ 0.2 - 50 RPM
- 100RPM
0 - 150RPM
200 RPM
"0 ·22l.!:::::==
2 .c5=
: =3!___3,J..5
- o--J.
4 _ _4.L.5_ _ j5 " 0 · 221.!:::=~2.~5~~3'--_.i_-_i_ _4.L.5, _ - J5
Propeller Diameter (m) Propeller Diameter (rn)
Number of Blades: 5 Number of Blades: 6

,: =c; ~s::~~ r;
O .Sr---,----,----==~==::-1
:---- 1

······~~.
g 0.4 CD l , , ] -~,,,'-~
.!!!
~
u
il} 0.2 - 50 RPM 0.2 - 50 RPM ·····························
- - 100RPM 100 RPM ~
0 - 150 RPM ···· ·····-··········-····· 0 - 150RPM ·········f··· · ······~·· ········· ~· ·········
200 RPM 200 RPM . . .
2.5 3 3.5 4 4.5 5 2.5 3 3.5 4 4.5 5
Propeller Diameter (m) Propeller Diameter (m)

Figure 3: Efficiency Plot

2.3.2 Blade Design

After the parameters of a propeller with a viable efficiency curve have been established,

the desired inputs are entered into the propeller design option, shown in Figure 4.

19
 ~ Nu rnb~tr of Bladu 1:1 flub Image Flag (Check for YES)
~ Propeller Speed (RPM)
~ Propeller Diameter (m) Meenlino Typo: Thickneo:o Form:
)NACAISA010

I .ttxllO R•quir•d Thnnl (N)

~ Ship Veloc ity (m/1)
~ Hub Olameter(m)
~ Number of Vorta11 Panels over the Rdius
~ Mu. Iteration• m Wake Alignment
~ Hub Vor1ea R8dius/Hub Rediu•
r-o-- Hub Unloading F.ctor: lPOptimum
r-o-- Tip UnloadiAg Fecfotr 1• Recluced Loading
~ S>Mrt Cancellation factor: 1• No Cancellation
~ Willer Ounsily (kllfrn"3)

~ Shaft Centerline Ofl:plh (m)

~ Inflow Variation (mil) Op1:1nProp

~ Ideal Angle of Atlac k (cfegreu )
~ Number or Pomta onr the Chord

Figure 4: Propeller Design Matlab Interface

Several additional inputs are supported in the blade design function.

• Shaft centerline depth: The unit of this parameter is meters (m). It presents the

depth when a propeller works.

• Inflow variation: It is required for the calculation of pitch angle variation, and the

unit is m/s [17).

• Ideal angle of attack: This parameter is to calculate the pitch angle, and the unit is

degree.

• The number of points over the chord: This parameter decides the resolution of a

propeller blade. Usually, 20 points can provide sufficient resolution [ 18).

In the program, two types of mean line, which is a line drawn midway between the upper

and lower surface, are available: the NACA (National Advisory Committee for

Aeronautics) a=0.8 and the parabolic meanline. There are three types of thickness forms:

NACA 65 AOl 0, elliptical, and parabolic. The thickness form of ACA 65 AOI 0 is

20
designed to obtain high lift coefficient and high speed. The NACA thickness form is

combied with the meanline a=0.8 in this OpenPVL code to construct propeller foil

sections. The meanline of a=0.8 indicates that the pressure distribution on 80% of the foil

chord is uniform [ 19]. The thickness form of elliptical is designed to reduce drag and

obtain a thin blade with necessary strength. The thickness form of parabolic combined

with the parabolic mean line is used to reduce resistance force and is applied for high-

speed applications. Nowadays, the commonly used foil sections are the NACA foil

sections, which includes some series of models, such as 4-digit series, 5-digit series, 16-

series, 6-series, 7-series [ 19]. Each series has its own advantages and disadvantages.

Table 2 displays the advantages and disadvantages of each NACA series model.

21
 Table 2: Advantages and Disadvantages of Each NACA Series [19]

Series Advantages Disadvantages

4-Digit Series Good stall characteristics; Low maximum lift

Roughness has little effect coefficient; high drag

5-Digit Series High maximum lift Poor stall characteristics;

coefficient; roughness has high drag

little effect

16-Series Avoid low pressure peaks; Low lift

low drag at high speed

6-Series High maximum lift High drag outside of the

coefficient; low drag in the operating conditions

operating conditions;

optimized for high speed

7-Series Low drag in the operating Low lift coefficient; high

conditions drag outside of the

operating conditions

The NACA 65 A01 0 foil section is the only NACA model in the OpenPVL code;

however, there is the potential for users to add more foil sections in the OpenPVL code in

the thickness form and meanline sections in the original OpenPVL code.

22
The parameters that are on the right side of the blade design screen are introduced a

below.

• f0/c and tole: fo/c is the maximum camber distribution. t0/c is the maximum

thickness distribution. The maximum camber fo and the maximum thickness to are

shown in Figure 5. cis the length of the nose-tail line, which is the dashed line in

the Figure 6.

• Skew: The unit of skew is degree. This parameter is applied to reduce the

propeller-induced unsteady forces [17].

• Xs/D: This parameter is the non-dimensional rake. It is applied to reduce the

propeller induced vibration [17].

/
J ~Iean Line
/
/
/

/ ------------------~
IMaximum T~lckness, 1 0 ] /
/ !Maximum- Camber, fa l

!Chord ! ------ ~

Figure 5. Foil Section Geometry [17]

After the calculation process, the graphical reports are created. In Figure 6, the upper left

comer shows the non-dimensional circulation vs. the radial position, and the upper right

shows the axial and tangential inflow velocities with the axial and tangential induced

23
velocities vs. the radial position. The lower left comer shows the undisturbed flow angle

and the hydrodynamic pitch angle vs. the radial position, and the lower right shows the

chord distribution vs. the radial position.

...W.!!.

J=0.750; Ct=0.994; Cq=0.346; Kt=0.220; Kq=0.038; '1=0.686

c 0.03, - - - - - , - - ---,- - - . . , . - ---,
..g : i : - VaNs
~ 0.026 ................c........... -~ ................ ;.. .. ........ . - - VWs
~ : : : - - Uo' Ns
i3 0.02 ................................ .;. ................ ;............ .. 0.5 ......... ""t ........
-.; : : : ----· Ut'Ns
5 0.016 ..............~ ................ ~ ... ......... ;.............. . ~ . ~.- . -j· - . . - -. j. -
' (i; : : '

~o....E1- oo.o.oo51 :::::::::::::J,_,::::::::::::::::r.,:_ ::::······--····]····--·····

....
..···
-------~----············
0 ---~---~----~---
: ""'l .,. ; ...
z . .
%.2 0.4 0.6 0.8 ' 0 ·n '=2-
-. --7
o_L.
4 ---=-'6 --0:::".8:---~
o.-=-
r/R r/R

so r---::---:---r==~~

50 ~ ..,,.......
'
-,
...l: ................l: .............!\-,- - - ' 1~
40 ..~~-~~~ ................ +.... .. ....... ~................
=: :,
-
:
'- ;'-....._ : e
0.15 ................~ ................ .:. ................ ;............ .
. . '

r~~
u 0.1 .... ········+·········· ·····+ ..
E:: 0.05 ..........

0.4 0.6 0.8 %.2 0.4 0.6 0.8

r/R r/R

Figure 6: Graphical Report

Figure 7 presents five propeller blade profiles in a two dimensional view. This figure also

displays the chord length, the pitch angle, the camber, and the thickness of the propeller

foil section. Figure 8 shows a three-dimensional propeller image. It provides an instant

graphical presentation of the propeller design showing the number ofblades selected and

a single cylindrical hub.

24
 2D Blade Image
v-~
- r/R = 0.20123

0.1 -
-
r/R = 0.29584
r/R = 0.50662
r/R = 0.75307
-··· •······ . . .... . . . . . [7/L j~A:
- r/R=0.94106
0 .05 ----··r ------------~------------· · !·--------- - -- ~-:--· ········---~- .......... .

0 -··· --·······r·· -------- ·:········· ········-··············· ......... ·····

-0.05 -·----'-··-·-··-····-i---······-···-i-- I. ---~>/··-------··-r··--------·--'--------------!·-----------------------

-0.1 ' : --···

·····;··-~-~.....

J / :
/ ----·--·-·-·--·

/ :
-0.25 -0.2 -0.15 -0.1 -0.05 0 0.05 0 .1 0.15
X (2D)[m)

Figure 7: 2D Propeller Blade Profile

ffe'dtf-lf-tlools~~~

_c.l ~~"~'- n 1Bl_:.o .~ --l~- o @l" c

30 Propeller Image

0 .5

I
g 0
N

-0.5

-1

0
-0.5
·1 -0.5
Y(3D)(m)
X(3D){m)

Figure 8: 3D Propeller Image

25
2.3.3 Propeller Geometry Development Through CAD

After the blade design is determined to be satisfactory, the program will create a text file

(OpenPVL_CADblade.txt) to export the propeller blade geometry. The file provides a

series of points that describe each foil section of the blade. This file was then used as an

input for the CAD software RHINO, which generates the points for the foil sections. The

propeller geometry is generated by creating closed splines for each sections and

c01mecting the sections to generate a blade. The hub design is the following step, and then

based on the designed number of blades to add all blades onto the hub [20]. After the

propeller geometry is generated, a file can be created by RHINO to fabricate a prototype

for testing by a rapid prototyping (RP) machine (.STL format) . The RP part can then be

tested. If the testing results are not satisfied, designers can go back to modify the

propeller geometry in RHINO or regenerate a new design in OpenPVL, and then

prototype and test again. These fabrication-testing-modification procedures are repeated

until a satisfactory propeller is created.

In 2007, D'Epagnier designed an AUV propeller using OpenPVL code, and the AUV

Propeller has the following characteristics [20] :

• The propeller has three blades.

• The propeller is operated on the vehicle at 120 RPM in order to reach a speed of

1.0 m/s.

• The diameter of the propeller is 0.6096 m.

• The required forward thrust of the propeller is 75 N.

26
 • The diameter ofthe hub is 0.12192 m.

• Inflow wake velocity variation is - 0.03 tnls.

After running the blade design function with the propeller parameters, the blade geometry

was created in RHINO by importing the OpenPVL_CADblade.txt file and then

manipulating the blade as described. Figure 9 shows the propeller in RHINO.

Figure 9. Propeller in Rhino [30]

D 'Epagnier proved that the actual FDM-printed blade geometry was the same as the

desired blade geometry that was determined in OpenPVL [20]. D'Epagnier used a milling

machine with a dial indicator to measure the propeller blade with three different r/R

values, 0.25, 0.70 and 0.80. Along each r/R value five evenly-spaced points were chosen

for testing. In Figure 10, the green lines are the desired blade geometry with the three

different r/R values of the AUV Propeller, and the triangular points are the tested points

27
of the actual 3D printed blade. Due to the difficult in pinpointing the edge of the blade

with a dial indicator and the effects of deflection, the measurements of the blade

geometry at the leading and trailing edge are somewhat less in aggreement. However, the

other points show good alignment with the desired results. This experiment was only

applied for geometry validation that the produced RP test propeller was the desired

propeller, which was designed in OpenPVL. The experiment didn't display the propeller's

hydrodynamic parameters.

n
Blade Cross-Section at t/R- 0.25 "' Actual Blade Geometry
... Desred Blade Geometry

·"'
i .. "'
I I I r I I I
~ 26 2... v 2.7.5 2.1 2 85 2.0 3.05
Along-Sp<Vl.X !itt)
Blade Cross-Section at 1/R- 0.70 .o. Actual Blade Geometry
... Desired Bk1de Geometry

H
<~ 122 "' u•
I
....I
....
I I
1.3
Along-Span,x(in.)
I
8.92 ....
I
:.

8.36

...
I ,_,.
I
..
I

n
Blade Cross-Section at r! R- 0.80 Actual Blade Geometry
Desired Blad9 Geometry
"' :.
->
...
0
C(
..
\j, J8
"' ..
I I
952 ....
I

Along-Span.x(in.,
I
9!6 ....
I
..
I

Figure 10: Propeller Geometry Validation [30]

2.4 Review of Computer-Aided Design (CAD)

Normally, propeller design codes are combined with Computer-Aided Design (CAD)

technology to complete the design . Computer-Aided Design (CAD) is defined as, the use

of computer technology for the design of objects. Started in the late 1980s, Computer-

Aided Design programs were used to design curves and figures in two-dimensional (2D)

space or curves, surface and solid in three-dimensional (3D) objects, thus beginning a

28
trend for many companies to reduce cost in drafting departments. As a general rule, one

CAD operator could replace at least three to five drafters who designs by hand. CAD

could be used in many applications, such as the automotive, shipbuilding, aerospace

industries, industrial and architectural design [21]. Due to fast paced development of

personal computers, a large number of CAD software packages have been created and

developed from 2D to 3D, from simple to complicated design and solid modeling, such a ,

AutoCAD, CADRA, MiniCAD, Euclid-IS, Pro/Engineer, SolidWorks, Solid Edge and so

on [22].

SolidWorks, a mid-price CAD software package, has a large number of customers

worldwide. In this thesis, SolidWorks is the CAD software used for propeller design and

simulation. This CAD software is based on the parasolids solid modeler and utilizes a

parametric feature-based approach to create models and assemblies for 3D mechanical

design and solid modeling. Started in 1995, SolidWorks already has many applications

and tools for mechanical design, such as drawing tools, design validation tools, product

data management tools, design communication and collaboration tools and CAD

productivity tools. SolidWorks has become a comprehensive mechanical CAE

(Computer-Aided Engineering) software, to aid in engineering tasks [23].

CosmosFloWorks is the Computational Fluid Dynamics (CFD) application in SolidWorks

and can be applied to fluid-flow simulation and thermal analysis. CosmosWorks is the

Finite Element Analysis (FEA) application that can be used for stress analysis.

29
2.5 Review of Computational Fluid Dynamics (CFD) and Finite Element Analysis

(FEA)

The technology of Computational Fluid Dynamics (CFD) as a design validation tool is

brought into SolidWorks to use numerical methods and algorithms to solve and analyze

problems related to fluid flows . CFO technology discretizes the spatial domain into small

cells to form a volume mesh or grid, and then applies a suitable algorithm to solve the

equations of motion. Treating a continuous fluid in a discretized format is the most

fundamental consideration in CFD [24].

The Finite Element Method (FEM) and the Finite Volume Method (FVM) are the two

main discretization methods widely used in the CFD field. The FEM uses standard

techniques for finding solutions of partial differential equations (POE) as well as integral

equations. The solution approach is based on eliminating the differential equation

completely or converting the PDE into an approximating system of ordinary differential

equations, which are then numerically integrated using standard techniques such as

Euler's method, Runge-Kutta and so on. FEM is mostly used in solving partial

differential equations over complicated geometry, such as cars, to increase prediction

accuracy in important areas [25].

Finite volume method (FVM) is a classical approach used most often in commercial and

research codes. The governing equations are solved on discrete control volumes. FVM

recasts the partial differential equation (POE) of the Navier-Stokes equation in the

conservative fonn and then discretizes the equations. This guarantees the conservation of

30
fluxes through a particular control volume. Another advantage of the finite volume

method (FVM) is that it is relatively easy to formulate for unstructured meshes [26].

FVM is often used in dynamic flow analysis, like marine propellers. In 2008, Pawel

Dymarski [27] published a paper that presented a computer program SOLAGA to

compute viscous flow around a ship propeller. The numerical model used for solving the

system of main equations is based on FVM. The solution domain is subdivided into a

finite number of control volumes, which are solved based on the integral from of the

conservation equations. In the results, the calculated pressure distribution over the blades

of the propeller is smooth, and the calculated propeller thrust and torque are in agreement

with the experimental results. This paper showed that FVM is applicable to flow dynamic

analysis for propellers. Presently, Finite Volume Method (FVM) is used in many

computational fluid dynamics packages, such as, CosmosFloWorks, which is the

SolidWorks integrated fluid simulation application.

SolidWorks also provides highly-advanced Finite Element Analysis (FEA) functions to

designers and engineers. FEA is a numerical technique to solve engineering analysis

problems for structural and field applications. The idea of FEA is to break a complicated

structure into small elements and each element is based on physical law to calculate

algebraic equations and solve engineering problems [28]. CosmosWorks is the

application package for FEA analysis. CosmosFloWorks can calculate the surface

pressure results of a propeller, and then transfer the results into CosmosWorks to do a

stress analysis. CosmosWorks use the pressure results as an input of the stress analysis,

31
and then to check the strength of the propeller design. The detailed steps of CFD and FEA

for a propeller design are listed in Chapter 3.

2.6 Review of Propeller Fabrication

Following the design validation through CAE, a real prototype part is needed to be

fabricated for testing. With the development of computer aided engineering (CAE)

software, some software applications are available for propeller simulation, which can

predict the results of a propeller test and be a convenient way for engineers to speed up

the design of the final propeller. Because simulation does not reflect all the aspects of the

real world accurately, simulation models are designed and used with the goal of

approximating the testing process, but, can't completely replace the testing process.

Physical prototype testing is used to increase to an acceptable level the confidence that

the simulation results are correct for the real component [29]. For this reason, propellers

need to be fabricated for testing after the simulation. There are generally three methods of

propeller fabrication: Casting and Computed Numerically Control (CNC) machining

technology are the traditional methods. In the late 1980s, rapid prototyping (RP) was

introduced and by the late 90's it was used as a low cost and fast process for physical

prototype fabrication.

2.6.1 Traditional Fabrication Processes

Casting is a manufacturing process which involves pouring liquid material into a mould

designed with the desired shape, and then allowing to solidify. The solidified component

32
known as a casting, is usually ejected or broken out of the mould and then put through a

finishing process [30]. Figure 11 is the flow chart of a standard casting process.

Build Pour Take Finishing

mould liquid Solidification out the machining
material part

Figure 11: Flow Chart of Casting Processes

As a traditional manufacturing process, casting consists of several major steps from

inception to completion of a product. These are: demand for a casting of specific shape

and size; production of drawings, patterns or prototypes; the application of simple

experienced-based rules to ensure good molten metal behavior. All these require some

engineering to produce a good casting and will often take several attempts before a

satisfactory result is obtained upon the development of a new product [31]. Due to these

requirements of casting, cost and process time are major problems for building test

prototypes. CNC machining technology is also commonly used to produce propellers. In

the CNC process, CAD files provide the input for computer aided manufacturing

programs, which are commonly used to extract computer file for a component and to

extract the commands that are loaded into the CNC machines for production. CNC

process can be significantly faster than casting but they are still fairly time consuming and

expensive. For example, the time to produce a propeller with 200mm diameter usually

takes more than two weeks, and the cost is around $2000, if the material is brass

(Technical Service, Memorial University). Compared with CNC technology, casting will

take a longer time and cost much more, because several processing technologies will be

used for finishing machining. However, casting has two advantages: casting can save

33
much more time than CNC for volume production; another is that compared with cutting

off material in CNC, casting can save material. However, both of the casting and CNC

technologies take a long processing time, which is inconvenient for a rapid propeller-

testing requirement. This is a real problem when the actual performance has yet to be

tested and the design is not entirely finalized. Rapid prototyping is a much faster and

cheaper fabrication technology for rapid product development.

2.6.2 Rapid Fabrication Processes

In the late 1980s, the first technique for rapid prototyping became available, and was used

to produce models and prototype parts. Rapid prototyping (RP), using additive techniques

for the automatic construction of physical objects directly from three dimentional CAD

data, can significantly reduce the time for the product development cycle and improve the

final quality of the designed product [32]. A large number ofprototyping technologies are

available in the marketplace, such as, selective laser sintering (SLS), fused deposition

modeling (FDM), stereolithography (SLA), Laminated object manufacturing (LOM),

Solid Ground Curing (SGC) and 3D printing (3DP) [33]. Each prototyping technology

uses specific materials and techniques to manufacture parts. Selecting the appropriate

teclmology for RP fabrication depends on the desired accuracy and material requirements

of the component being produced. Table 3 shows the material for each prototyping

technology.

34
 Table 3: Prototyping Technology Materials [34]

Prototyping technologies Materials

Selective laser sintering (SLS) Polycarbonate; Nylon; Glass filled nylon;

Copper-impregnated nylon; Flexible

rubber; Steel; Silica based sand; Zircon

based sand; Investment casting wax

Fused deposition modeling (FDM) Acrylonitrile Butadiene Styrene (ABS);

Medical grade ABS ; Methyl methacrylate

ABS; Polycarbonate plastic; Investment

casting wax;

Stereolithography (SLA) Acrylin resin; Bi-colour acrylic resin;

Epoxy resin; High temperature epoxy

resin; Flexible epoxy resin

Laminated object manufacturing (LOM) Adhesive backed paper; Adhensive

backed polymer; Adhensive backed glass

fibre

Solid Ground Curing (SGC) Photoreactive resin

3D printing (3DP) Water based liquid binder on cellulose

starch powder formulation

Fused deposition modeling (FDM) [28] was developed by S. Scott Crump in 1990. The

principle ofFDM like all RP technologies is to lay down material in layers. An extrusion

35
nozzle is heated to melt the material and can be moved in both horizontal and vertical

directions by numerical controlled mechanism, directly controlled by a computer-aided

manufacturing (CAM) software package. The main materials for FDM include

Acrylonitrile Butadiene Styrene (ABS); Medical grade ABS; Methyl methacrylate ABS ;

Polycarbonate; Investment casting wax [34] . The process of the FDM machine is fast and

relatively low cost. In 2007, Nadooshan [35] used FDM technology to construct a wind

tunnel model with Polycarbonate plastic material, which is an actual impact-resistant

industrial-grade thermoplastic and is structurally strong. Traditionally, wind tunnel

models are made of metal and are very expensive. FDM was used as a way to reduce time

and cost. Figure 12 displays the wind tunnel model constructed by FDM.

Figure 12: Wind Tunnel Model Constructed by FDM

This wind tunnel model was tested by engineers to compare with the real values of this

wind tunnel model with metal material. The most purpose of this test in the wind tunnel is

forces and moments. The results displayed that the accuracy of the data is lower than that

of a metal model due to surface finish and dimensional tolerances, but the FDM model is

36
quite accurate for a testing level. This FDM model cost about $650 and took 4 days to

construct, while the metal model cost about $1300 and took a month to design and

fabricate. The conclusion was that FDM technology is a timely and cost effective way of

producing test parts.

In this wind tunnel model, the material needs to be strong enough to sustain the air force,

which is on the nose cone and the edges of the wing tails. However, forces on a marine

propeller are around all blades and nose cone and the fluid is much more dense. If a

propeller needs to provide a large thrust for a vessel, this polycarbonate plastic material

may not be strong enough. Selective Laser Sintering (SLS) is a better choice to produce a

propeller where higher strength is required. SLS [34) is a rapid prototyping technology

that uses a high power laser to fuse small particles of plastic, metal, ceramic or glass

powers into a mass to build a desired 3-dimensional object. The laser selectively fuses

material by scanning cross-sections, which will be used as the solid part of the object. The

scanned parts are melted and become a solid. After each cross-section is scanned, the

platen on which the object is built is moved down by one layer thickness, a new layer of

material is applied on top of the solidified layer and the process is repeated until the

object is completed. As shown in table 2, the SLS technology has a wide range of

available materials. Using SLS technology, objects can be manufactured from high

strength materials, such as uncoated or polymer-coated steel powders, which are

unavailable for other technlogies. The layer thickness of SLS is 0.001-0.004 inch,and

laser diameter is 0.004-0.02 inch [36). With the small heating diameter, thin process layer

and high strength of material, propellers can be produced by SLS with a good accuracy,

37
surface finish and strong solid body. In this thesis, the availble prototyping machine was

an FDM 2000 which was adequate for small testing propellers. A small tip can be used to

make a smooth surface and the material ofPolycarbonate Plastic (PC) can be used to

make a strong propeller.

2.7 Review of Propeller Testing

After a propeller is fabricated, the physical testing procedure is used to test the propeller

performance. Thrust is a very significant parameter for a propeller along with torque, o

the testing device is set up to test the propeller thrust. Figure 13 is a sketch of the

equipment setting for propeller testing.

D'

-'11.

arnage

Water

Thruste~
mount ~ropeller

Tank

Figure 13 : Sketch of a Propeller Testing Set-up

38
In the experiment for the propeller thrust test, a tow-tank is set up to provide the water

domain. The tested propeller is mounted on a thruster mount. Figure 14 shows an

example equipment set-ups of the thruster mount. The thruster mount is connected with

the tow-tank carriage, which, working as the speed supply, is controlled to move with

specific speeds to simulate ship motion. The equipment of a tow-tank carriage is shown in

Figure 15. A load cell is attached in compression to the top of the thruster mount, and

connects with a computer. In the computer, there is a program that is used to convert the

load cell signals to calculate propeller thrust.

Figure 14: Thruster Mount

39
 Figure 15: Tow-tank Carriage

Sometimes, the propeller size and thrust value are very large. Due to the limitation of cost,

tank size and testing range of dynamometer, propellers are usually tested from a smaller

size model first. The laws of similarity are used to provide the conditions, under which a

model must operate, so that its performance will reflect the performance of the prototype

[ 15]. The first condition is the geometrical similarity, which requires that the model' s

geometry is similar to the full-size propeller. Because of this, the ratio of every linear

dimension of the full-size propeller is a constant ratio to the corresponding dimension of

the model. If the diameter of a propeller is 5m, and the diameter of the model is 1m, the

scale ratio A-=5/ 1=5, and this ratio should be held constant for all linear dimensions, such

as the hub diameter, the chord lengths and blade thickness. The second condition is the

kinematic similarity, which requires that the ratio of any velocity in the flow field of the

40
full-size propeller to the corresponding velocity in the model is constant [15]. It reflects

by equations as below.

[ 15]

S refers to the ship and M refers to the model. VA is the speed of advance, and unit is

meter per second (m/s). n is propeller revolution rate, and unit is revolution per minute

(rpm). Dis the propeller's diameter, and unit is meter (m). J=V A/nD is advance

coefficient.

The third condition is the kinetic similarity, which requires that the ratio of the various

forces acting on the full-size propeller is equal to the corresponding ratios in the model.

This means the Froude number (Fn), Reynolds number (Rn) and Euler number (En) of the

full-size propeller is equal to the model ' s.

R = VAD p
II > E, =1 [ 15]
v - pV 2
2 A
g is acceleration due to gravity, g:::;9.8lm/s2 . - )li p is the kinematic viscosity of the fluid.

pis mass density ofwater, p= l000kg/rn 3 . pis the pressure associated with the propeller

and the flow around it. ll is the coefficient of dynamic viscosity.

Furthermore, thrust coefficient (KT) and torque coefficient (KQ) of the full-size propeller

is equal to the model's.

K T K = Q [1 5]
T = pn2D4' Q pn2D s
Tis thrust, and unit is Newton. Q is torque, and unit is Newton meter (Nm).

41
For given values of J, Fn, Rn, En, the values ofKT and Ko are the same for the full-si ze
propeller and it is geometrically similar model. All the above relations are used to

calculate the thrust value for a propeller model test. If the tested value is close to the

calculated value, it means the propeller has a desired performance as designed.

42
Chapter 3

Marine Propeller Design and Simulation in

SolidWorks
Although the OpenPVL code has been developed extensively, there is still room for

enhancement. Before making a prototype, simulation work is commonly done to predict

the performance of the designed propeller. Based on simulation results, appropriate

modifications will be applied, yet RHINO does not directly support simulation work. This

simulation and analysis work has to be done by another software, which is not ideal. It is

advisable to generate the propeller geometry in a CAD program, which can also directl y

support the simulation work. Solidworks, as an integrated CAD software package, is not

only able to generate propeller geometry, but can also simulate the propeller fluid

dynamics using the application CosmosFloWorks (CFD) and check the strength of the

propeller design by using CosmosWorks (FEA). Updating the OpenPVL code and

transferring the propeller geometry data into Solidworks became the continued issue. The

author expands the application of Open_PVL code by creating OpenPVL_ SW to generate

43
propeller geometry for import into the SolidWorks software to initiate the simulation

work.

In Chapter 3, the use of the OpenPVL_SW code for marine propeller design and the use

of SolidWorks to generate propeller blade geometry and validate it through simulation

will be discussed. In OpenPVL_ SW, the parametric analysis function allows the user to

combine several propeller parameters to optimize the propeller design, and then to use the

optimized parameters to design a propeller blade. An output file

OpenProp_ Solidworks.txt from the propeller design function will be used as an input to

generate the propeller blade structure in Solidworks. After a propeller blade is created, a

hub will be designed. Based on the geometry of the blade and hub, other blades can be

generated by the Circular Pattern function in SolidWorks. After a propeller is generated,

CosmosFloworks and CosmosWorks will be utilized for propeller simulation.

3.1 Open_ PVL and OpenPVL_SW codes

OpenPVL_SW is an extended version ofOpen_PVL to be able to create propeller

geometry in SolidWorks, which can simulate the propeller working process using the

integrated application CosmosFloWorks. The OpenPVL_SW code uses the same

propeller analysis and design method as Open_PVL and is extended to generate a

propeller blade geometry output file for SolidWorks. Compared with the Open_PVL code,

the OpenPVL_SW code has two main advantages: The first, Open_PVL code generates a

propeller blade geometry file for RHINO, which can not do the simulation work for a

propeller. Thus, if engineers need to do a simulation to confirm the propeller design is

44
good enough for a real testing, they have to transfer the propeller geometry into other

simulation software. This is a waste of time during a large number of simulations and is

inconvenient for engineers. The OpenPVL_SW code can generate a propeller blade

geometry file for SolidWorks, which can use the integrated simulation application

CosmosFloWorks to do a CFD simulation. This can save a lot of time and is much more

convenient for engineers. Furthermore, the pressure output from the CFD simulation can

be input into CosmosWorks that can be used to check the strength of the propeller design .

The second, the propeller blade geometry file created by Open_PVL is the propeller foil

sections points in RHINO. During each cycle at propeller design, engineers have to use

several commands to create the propeller blade geometry and then design a full propeller.

This is also inconvenient for engineers. The propeller geometry file created by

OpenPVL_ SW can automatically create a propeller blade geometry without any manual

commands. This is much more convenient for engineers than Open_PVL.

As introduced in Chapter 2, the processes of propeller design by Open_PVL includes:

parameter analysis, design of a propeller blade, generation of the propeller blade

geometry in RHINO, hub design, adding other blades to complete the propeller, and

fabrication of the prototype propeller. Because OpenPVL_ SW code can create a propeller

in SolidWorks, the CosmosFloWorks and CosmosWorks applications can be used for a

propeller simulation to provide input into the design process prior to prototyping. Table 4

is the processes of propeller design and fabrication by Open_PVL and OpenPVL_SW.

45
 Table 4: Processes of Propeller Design by Open_PVL and OpenPVL_SW

Processes of Propeller design by Processes of Propeller design by

Open_PVL OpenPVL_SW

Parameter analysis

Selection of parameters

Design of propeller blade

Export blade section point data

Generate a propeller blade geometry in Generate a propeller blade geometry in

RHINO SolidWorks

Design a hub Design a hub

Add other blades to complete the Add other blades to complete the

propeller propeller

CFD by the SolidWorks integrated

Rapid Prototyping
application CosmosFlo Works

FEA by CosmosWorks to check the

strength of the propeller design

Adjust design as required

Rapid Prototyping

46
Using OpenPVL_SW, a propeller design can be validated by the simulation work in

CosmosFloWorks and CosmosWorks. The simulation work is quite convenient for

propeller engineers to design and validate a propeller before fabrication.

3.2 Using Solidworks for Propeller Blade Geometry

A propeller can be designed by the parametric analysis and propeller design functions of

OpenPVL. The OpenPVL_ SW code was modified to create an output of

OpenProp_Solidworks.txt file for generating a propeller blade in SolidWorks. After

running the propeller design option, the file Open Prop_Solid works.txt was created. Codes

for generating a propeller blade is recorded in the .txt format file. The .txt file records the

SolidWorks macro commands to generate a propeller blade. SolidWorks macro is a series

of commands and actions that can be stored and run within SolidWorks to automatically

draw geometry without manual working. Users can paste the codes in the window of

Solidworks macro to generate a propeller blade. The file has three main functions: to

generate foil section points for the propeller and connect each point to generate propeller

foil sections; to connect each leading edge point of the foil section to create a reference

line for the blade surface; to use SolidWorks "loft" function to generate the propeller

blade based on the foil sections and the reference line. The flow chart shows how to

generate a propeller blade in Figure 16.

47
 Connect Generate Generate a
Generate points to leading propeller
foil section generate edge Line blade by
points foil of the "Loft"
sections propeller function

Figure 16: Flow Chart of Generating a Propeller Blade

3.2.1 Propeller Foil Sections

The foil section points are three dimensional values (x,y,z), therefore the SolidWorks

function "3D sketch" is used to read the points information into SolidWorks. As

referenced in Chung' s paper [18], 40 points on a propeller foil section can provide

sufficient resolution in geometry. OpenPVL_ SW defines that each foil section has 40

points, which are in the order of 1-40. If users want to change the amount of foil section

points, they can change the point amount in the OpenPVL_ SW code, and confirm that the

point amount is the same as the amount displayed in the propeller design function of

Matlab graphical user interface (See in Chapter 2). The foil section point amount is

defined by the commands ofReDim swSketchPt 0 in the OpenPVL_ SW code. The

number in the parenthesis mark is the number of foil section points. A Large number of

points can provide a high resolution in propeller foil section geometry, however, too

many points will waste more time on unnecessary calculations. Figure 17 shows the

points on a propeller foil section.

48
 . 1 :w
6 '9· . ~ 17'8 !c- ~ . 1 "'S
I> ao u:~ D .s~

Upper Chord

3
18
1(40) • Leadmg Edge • 1.9 :w
Trailing Edge • •
39.
22 21
38 23

Lower Chord
l

Figure 17: Foil Section Points

A curve passes through all the 40 points to generate a propeller foil section. The curve

starts from point number 1, which is on the leading edge of a propeller blade. The curve

will pass through all the upper chord points in an orderly way to point number 20, which

is located on trailing edge of the propeller blade. After that, the curve will orderly pas

through point number 21 to point number 40 (coincident with point number 1), which are

on the lower chord of the propeller blade. As referenced in Karim 's paper [37], B-spline

is used as the curve to generate the propeller foil sections. The shape of the B-spline

curve depends on the directions of tangent vector. B-spline can have as few as two points

and can specify tangency at the end points. To achieve higher accuracy, a large number of

points are needed for the approximation. In the OpenPVL_ SW, 40 points are used in B-

spline to provide a smooth and accurate curve for the propeller foil sections. A closed B-

spline starts from the point number 1 and end at the point number 40, so the B-splinc has

49
the same tangency value at the leading edge point. Thus, a closed B-spline can provide a

smooth curve at the leading edge point. In order to confirm that a propeller foil section is

a closed curve, OpenPVL_ SW defines that point number 1 and point number 40 have the

same three dimensional values. The foil section is shown as Figure 18.

~"" "" ,.. _ . 1""' , _ . . - 1100

D-a.Hl~'l!! ~· ..., . . ~'it< S I!!J !f3. 5liiJ:-J1.LI
!> N ()9 Cl (;
~m
,._.,,..,
. . til
f)Ct.ruclt... lltl'
o;.
R....UV.... ,~ "•
C3i a
5Md.8... t.ofttdl...

- Figure 18: A Propeller Foil Section

OpenPVL_SW generates 21 foil sections to construct a propeller blade. These foil

sections are defined in the order of 0-20 from blade root to tip. OpenPVL_ SW generates

the 21 foil sections in the order of0-20. These 21 foil sections are shown in Figure 19.

50
,flo (d< - """' !Pdo _ .. ,..,.,.. ""'
g)~

D ~' 1.~ ~ ':'!! .::, 19 • 11 9 , e !lf.l z. . ~ ·:rl .

I> 19 u Q1 (j CSI t
tll .
P:Nto.M
'll!
Sbtch
. lila
[)«)\ode
<!>
Pwf:#loe:,,

lJ
+•
I

a
ilfl 19

Tip 20 1\

Figure 19: Overview of Propeller Foil Sections

Figure 20 is the side view of these 21 foil sections that shows the overall propeller

geometry. These foil sections need to be linked by lofting a smooth surface to generate a

propeller blade.

51
 ~ flo td< "'" 11W«t I"" ,._,. - ,..
! ~;lriiJI!'(tl ~ 19 · . ~11. 8 lt:.l !c·!ll JJ· ~~ ~ ~
<> DO LJQI (j l l!l t
ulm ""( ~ . lij
.ul ~·~~"" r Sbtch £xtn.n....

Figure 20: Sideview of Propeller Foil Sections

Propeller designers can add more types of foil sections in the thickness form and

meanline sections of the original OpenPVL code. Every foil section is described by

meanline and thickness form information. Meanlines are described by two sets of factors,

the "foe" and the "dfdxN". The code ofNACA a=0.8 mean line is shown as below. The

factor "foe" is the ratio of camber to chord. The factor "dfdxN" is the slope of camber

line. The information of these two factors can be found in NACA reports. If propeller

designers want to use another mean line in the OpenPVL code, they can change the

numbers that are in factors "foe" and "dfdxN" to reflect the characteristics of the new

mean line.

52
Standard NA CA a=0.8 meanline factors:

foe = [0 .287 .404 .616 1.077 1.841 2.483 3.043 3.985 4. 748 5.367 5.863 6.248 6.528

6.709 6.79 6.77 6.644 6.405 6.037 5.514 4.7713.683 2.4351.163 0]./100

dfdxN = [.48535 .44925.40359 .34104.2 7718. 23868 .21050 .16892 .13 734 .11101

.08775 .06634.04601 .02613 .00620-.01433 -.03611 -.06010 -.08790 -. 12311

-.18412-.23921 -.25583-.24904 -.20385]

The NACA 65A010 thickness form is shown as below in OpenPVL code. The factor

"toe" is the ratio of thickness to chord. In the original OpenPVL code, toe_ 65 is named as

the ration of thickness to chord for the NACA 65A010. Users can rename the toe 65 for

convince. For example, toc_63 is used to define the ration of thickness to chord for

NACA 63A006. Propeller designers can change the thickness form factor to use another

thickness form other than NACA 65A01 0, if desired.

NACA 65A 010 thickness form

toc_65 = [0. 765 .928 1.183 1.623 2.182 2.65 3.04 3.658 4.127 4.483 4. 742 4.912

4.995 4.983 4.863 4.632 4.304 3.899 3.432 2.912 2. 352 1. 771 1.188. 604

.021].1100

Figure 21 displays the foil section geometry using thickness fonn ACA 65A01 0 and

a=0.8 meanline at the r/R=0.35.

53
~~mmE3~a. ................~~~==~~----~--~~~~-- ~~
~"' t.a r- ~ro-t lP<>k ,.._.. - tlob .....w
D Willi!~ · '9· · ~II> e ll:l !D· !>I iD · ? 'lS
~ E> uo CJIJI Cl r!l ,
]~ ~ • ..,,! . .....!.. . ...!. . ~ . ~0 . Ldt9a... F~

Figure 21: Foil Section using NACA 65A010 a=0.8 at the r/R=0.35

A foil section with a different thickness form (NACA 63A006) and meanline (a=0.8

modified) is provided as an example of how to change the geometry of foil sections. The

information about the NACA 63A006 and mean line a=0.8 (modified) was found in

NACA reports, and is shown in table 5.

54
 Table 5: The information ofNACA 63A006 and mean line a=0.8 (modified) [19]

toe 63 = [0 .495.595 .754 1.045 1.447 1.747 1.989 2.362 2.631 2.820

NACA 2.942 2.996 2.985 2.914 2.788 2.613 2.396 2.143 1.859 1.556 1.248

63A006 .939 .630 .322 .013]./100

foe = [0 .281 .396 .603 1.055 1.803 2.432 2.981 3.903 4.651 5.257

5.742 6.120 6.394 6.571 6.651 6.631 6.508 6.274 5.913 5.401 4.673

Mean Line 3.607 2.452 1.226 0] ./100

a=0.8 dfdxN = [.47539 .44004.39531 .33404.27149.23378.20618 .16546

(Modified) .13452 .I 0873 .08595 .06498 .04507 .02559 .00607 -.01404 -.03537-

.05887-.08610-.12058-.18034-.23430-.24521 -.24521 -.24521]

Figure 22 shows the foil section geometry using NACA 63A006 thickness form and

a=0.8(modified) meanline at r/R=0.35.

55
Pleal:eJekolrt•far:eOfasurfacetoolfsd.

Figure 22: Foil Section Using NACA 63A006 a=0.8(modified) at the r/R=0.35

The method for incorporating a change the thickness fonn and meanline is to change the

parameters of "foe", "dfdxN" and "toe" in the OpenPVL code as demonstrated.

3.2.2 Propeller Blade Surface Development

After the foil sections are created, a method is needed to connect them together to

generate a derived propeller blade. In OpenPVL_SW, a leading edge line is created on the

tip of the leading edge, which is shown as Figure 23. The leading edge line is created by

connecting all the point number 1(40) of the 21 foil sections. All the foil sections follow

the leading edge line to construct the propeller blade. The leading edge line is shown in

Figure 23.

56
 <:!l>oo '"' I'M - """' _.... - t>«>
Cl. 1<1 ~~ ~ '9 · ll VI ! .:;) ~ .~

u._,
e

Figure 23: Propeller Foil Sections and the Leading Edge Line

After the foil sections and the leading edge line are created, the SolidWorks " Loft"

function will be used to generate the propeller blade geometry. The loft function is used

to create a feature by following transitions between profiles in SolidWorks. Usually, this

function needs multiple sketches and a reference line to create the feature. Following the

leading edge reference line, the loft function connects all foil sections to generate the

propeller blade. OpenPVL_ SW code defines that the foil sections are lofted in the order

of 0-20, which is shown as Figure 19. The propeller blade is shown in Figure 24.

57
 <:ij,flotd<,....,..,l. . _ _ ti'l>

~ O ~ IdP\'9 ~ '9· . ~ m e li:l !e ·liliD· • J ~~ ~ • c+ e»- i!il!'j)CliCJj CJI!j o.!<>

] 1> ue tJ [][)' G!J~ 1!!- J

1l; ~~.:1· ,.!, . I,..,!!_... ....! . ,.._., ~... "''~'.. ~ ..~ ,. "' .1t.l!L oaf . . .!,

,.,\

Figure 24: A Propeller Blade

After a propeller blade is created, a hub needs to be designed based on the hub parameter

set out in the propeller design option of the OpenPVL_ SW code. Usually, a nose cone is

used on a hub to reduce the inflow effect when a propeller is rotating. Figure 25 shows a

propeller blade, a hub with the diameter of 0.2m and a nose cone with 0.05m.

58
 Figure 25: A Propeller Blade with a Hub and a Nose Cone

The following steps are used to generate other blades manually to complete the propeller

geometry. The axis of the hub is set up as a reference line, and SolidWorks "Circular

Pattern" function needs to be used to generate other blades based on the reference line.

Figure 26 is an example of propeller geometry with three blades.

59
 Reference line
.............
L.l .J

Figure 26: A Propeller with Three Blades

3.3 Simulations using CosmosFloWorks

Once the propeller geometry is fully described and satisfactory, CFD simulation work can

begin using the SohdWorks application "CosmosFloWorks". CosmosFloWorks is an

easy-to-use fluid-flow simulation and thermal analysis program that is fully embedded in

Sohdworks. CosmosFloWorks can be used to simulate thrust, which is one of the most

important parameter in a propeller design. Based on real experimental settings that were

introduced in chapter 2, a simulation structure is created in CosmosFloWorks. Because

CosmosFloWorks doesn't allow to set up the velocity parameter for a solid part, it cannot

simulate the motion of the carriage as described in chapter 2. Instead, the fluid velocity is

controlled in order to simulate the carriage motion. For example, if the test carriage

moves at 2 m/s forwards, CosmosFloworks can set up that the propeller is fixed with the

60
fluid moving at 2 m/s backwards. The resolution of the automatic meshing is controlled

through settings in CosmosFloWorks. The resolution level can be set to various levels

with finer and finer granularity. The software does not permit detailed configuration of

the mesh around particular geometric features. A higher level of resolution breaks the

simulation domain into more small elements and can provide much more accurate results

but it also requires significantly more computational effort. Determination of the exact

flows, stresses and deflections in areas such as the trailing edge of the blades (where there

are very fine geometric features) would benefit from an ability to adjust the mesh size

locally. The simulation settings and assumptions are provided in the following lists:

• Analysis type is external for propeller simulation. (ie. Fuild flows outside of

propeller)

• Water is used as the fluid domain.

• The only considered physical features are gravity and rotation, acceleration of

gravity g=9.81m/s2

• The temperature of water is 293 .2 K.

• The air pressure is 101325 Pa.

• Pressure potential is considered.

• Assume no cavitation is in the simulation

• Assume adiabatic wall is used .

The simulation structure is shown in Figure 27. This simulation model provides the water

domain, which serves the same function as the tow-tank in the real experiment.

61
 ~ ,_ t<lt !1o- 11-1 I"" I'll" ""- t!ol>
] O~,j f.I P 'I!I I? · llltd l ll;J :D·!i>l i!J · H 'lS I I C+ CJ· I @ BJBJ CJ ~ 0111 • .tiP
".. NuCi!TJ o 111 .. J r>~ tt !?.i 1'l. i11<l ll:!IW~ I 111 ' \1) ~ , :" i'\ 1\,

1 ~-:T ..!. · ...!. . ...!. . ...1.....!! . ~... '"S. I~ ~. ~ J! "'~' /....'!.... ....!!L oaf... ..~
1

,.::::::_
7
x_ Y

"
~

- Figure 27: Simulation Domain of Propeller

If users desire to model a specific test region and know the real testing tank size, they can

set up the simulation domain size the same as the actual one. Ifusers don't know the real

testing domain size, they need to define the domain size based on several simulation

results. When the propeller is rotating in the simulation domain, water is moving around

the propeller. If the size of simulation domain is small, the simulation programs will be

impacted by the flow effect at the edge of simulation domain. These impacts will cause

calculation errors for the propeller thrust prediction. Larger domain size is better for

simulation result accuracy, however; too large of a domain will increase the number of

simulation calculations, which can waste a lot of time. The guideline for simulation

domain design is to reduce the flow effect as low as possible at the domain edge. First,

the depth and width of a simulation domain are set up based on the propeller working

depth. If the propeller is designed to work at 3m below water, the depth and width of the

62
simulation domain are set to 6m and 6m. The propeller is in the center of the domain. The

second is the setting of domain length. In real propeller testing experience, 30m is a good

starting setting for a propeller simulation. The effect of moving water at the edge of

simulation domain can be monitored by water pressure. Figure 28 displays pressure

results of a propeller simulation with domain size of 6mx6mx30m (widthxdepthxlength).

The diameter ofthe propeller is 0.6m and rotation speed is 120 RPM.

Figure 28: Pressure Results of a Propeller Simulation with Domain Size 6mx6mx30m

(widthxdepthxlength)

In figure 28, most of pressure effect is within the simulation domain. That means the size

of simulation domain is appropriate for this propeller. Ifthe same simulation is run with

domain reduced to 4mx4mx20m (widthxdepthxlength), the results are shown in Figure

29.

63
- Figure 29: Pressure Results of a Propeller Simulation with Domain Size 4mx4mx20m

(widthxdepthxlength)

In this reduced simulation domain, there is much more pressure effect on the domain edge.

That means this domain size is not big enough for this propeller simulation. These

increased pressure effects will cause an error of propeller thrust prediction. Selecting a

simulation domain with little water pressure at the edge is a good choice for a propeller

simulation.

The simulation model provides water flow with a specific velocity to simulate ship

motion, which is the same function as the tow-tank carriage. This model also provides

propeller geometry. But, the model shown in figure 27 is not the same as a real

experiment. The propeller geometry is enlarged in figure 30, so the difference is shown

clearly.

64
 t.J- t*r i." - l-- CZ,. '\'"-WN ~ .tJtCI
u 'lri "'=IJ r.o g. 8r lf
., }Ill",, II'. 11:~

lil lil.l ~ :::: C> I& 1 '

1W tt•t ~,...._ ~P. ~ • ,.,, ,........,

Flow direction

Force on the hub

• •
Figure 30: Side View of a propeller

A hub is in the center of a propeller. There are two sides of a hub: the dome side is to

reduce the force from fluid; the flat side is to fix a propeller on a boat. As shown in Figure

14, the flat side of a hub is connected with a carriage and a thruster mount, so there is no

fluid force on this side. From the simulation model as shown in Figure 30, there is force

on the flat side of a hub when fluid is moving. This force will result in a discrepancy

between the simulation and real experimental results. A method to eliminate this force

will be introduced in the simulation model to get a more accurate result. Because all the

simulation parts in CosmosFloWorks have to be in the simulation domain, a design of

extending the flat hub side outside of the domain to eliminate the force on the flat side

doesn't work. Finally, the propeller simulation model is separated into two steps. First, a

simulation is run as the propeller structure in figure 30. Second, a simulation is run for the

propeller without blades, which is shown in figure 31 .

65
<:!i>E"~!lowii"MI""'Aow""'-t!OP
D~ ld 1118 '$ - 19 · -. ~ 'ih 8 ll!J
=-=- :c-. !>1
--: [ J::-.-u ;. t ~~ 1 -· 0 ~@ Bl Bl CJrJ I CJ Q ~ _,. ti>
I> IIO LJO!I CJ Clil l · I> @ .~ J f'l.fla'J l;l?,io ~~ '-'' 1m '"'• ~ ~ /* ~ 1\,1'\, t:f~ ...

11~ ~~" • .J:. · ,,!.. . ...!.. •,J:,. . ~. ~... ldt~o... l '~ ~- ~ ! c.~ 't. . l u..!,IL "'2 . . ..."! !.,.! . ·
....

n """''"' f

Figure 31: A Simulation Model Without Blades

The direction of thrust and force on the hub is opposite, so the force on the hub reduces

the simulation result for the thrust. The final simulation value of propeller thrust is the

simulated thrust value from CosmosFloWorks adds the force on the hub.

3.4 Simulations using CosmosWorks

CosmosWorks is a design analysis application, which is integrated in SolidWorks.

CosmosWorks uses the Finite Element Analysis (FEA) to simulate the working

conditions of designs and predict the behavior. With the fast solvers, CosmosWorks can

not only quickly analyze the designs and search for the optimum solution for designers,

but also shortens time to market by testing the designs on the computer instead of

expensive and time-consuming field tests.

66
CosmosWorks can be used to check the strength of the propeller design. In

CosmosFloWorks, the propeller is considered as a solid part without deformation;

however, the propeller will be undergoing stress when it is rotating. FEA analysis is used

to check that the propeller is strong enough when it is rotating. CosmosFloWorks can

provide the resultant fluid pressure that is acting on the propeller blades. This pressure

difference is the source ofbending force, which results the stress that could cause a

propeller to be destroyed. CosmosWorks can be used to analyze propeller's strength by

FEA. CosmosWorks uses the fluid pressure as an input parameter to calculate the stress

within the propeller and determine the Factor of Safety (FOS). The value of FOS is

calculated by limit yield stress over each stress value. The acceptable FOS value for a

propeller is not less than 1.5 [38]. Instead of 1.5, some designers use 2 to make a higher

strength propeller. The FEA analysis can help engineers to choose a good material and

hub design to make sure the propeller design is strong enough. The steps of FEA analysis

are shown as below:

• Use CosmosFloWorks to calculate the surface pressure of propeller blades

• Transfer the CosmosFloWorks results to CosmosWorks in the Tools option

• Create a static study in CosmosWorks

• Load the CosmosFloWorks results file (.fld) into the static study

• Set the restraint for the propeller. The bottom of the hub is fixed.

• Select propeller material

67
 • Set contact and centrifugal features for the propeller. Contact: bonded between the

hub and propeller blades; centrifugal: it is used to simulate the rotation ofthc

propeller, the unit is rad/s.

• Mesh and run the FEA

• Plot the Stress distribution and FOS results

• IfFOS ~ 1.5, the propeller is strong enough; ifFOS < 1.5, the propeller has a weak

strength, designers can change the propeller geometly to increase the strength or

use much more strong material for the propeller.

For example, Figure 32 displays the surface pressure of a three-blade propeller with 120

RPM working in 5 meters depth of water. The material of this propeller is Titanium Alloy

with 1.03 x 109 N/m2 yield strength as an example. In the propeller manufacturing, there

are several rules to select the material for a propeller [ 15]: the processes to fabricate a

propeller are casting and machining, so the propeller material must be amenable to these

processes; the propeller material should have a high strength and toughness, and the

fatigue strength is important as well ; resistance to conosion in water is desirable quality

in a propeller material; the propeller can be easily repaired if it is damaged; the cost of the

propeller material is also a consideration. In this thesis, the materi al used in FEA onl y i

an example without detailed considerations, and to display how the Cosmos Works works.

In a real propeller deisgn, the selected material should select based on the rules.

68
1Custom H

Figure 32: Surface Total Pressure Distribution of a 3-Blade Propeller

Figure 32 displays the total surface pressure, which includes static pressure and dynamic

pressure. Static pressure is affected by water depth, and the equation is p= pgh. Propeller

thrust is provided by the difference pressure between both sides of blades. Static pressure

on the two sides of blades is the same, so it doesn' t provide thrust. Dynamic pressure is

the difference pressure on propeller blades to provide thrust. Figure 33 displays the

dynamic pressure distribution of the propeller.

69
 Figure 33: Dynamic Pressure Distribution of a 3-Blade Propeller

The total pressure distribution is the dynamic pressure (symmetrical distribution)

combined with static pressure (changed by water depth). CosmosWorks use the surface

pressure result as an FEA input parameter to analysis the propeller' s strength. The

pressure result needs to be loaded in the flow condition frrst, and then bonded and

centrifugal force features are considered in this FEA. After finishing settings, run the

FEA simulation and Figure 34 displays the stress results of the propeller.

70
 e · IIJ· ~
~'111 1 li So.-!11 ~ ill II ' Ill l'l. llli!l ll ll . ,. " • D • "$ "
..!. . ..!.. ..!.. ......
~oa
..!.. ~ .! ~
at ....._....,... .s a.S. ! ! .! J! o!. .!. ...!.. .
.! ...... .-..n~ ~
~ .

EducatlonaJ Veralon. For lnttructtonll Un Onty

Figure34: Stress Distribution of the 3-Blade Propeller

In Figure 34, the areas near the hub has the higher stress, and these areas are needed to be

seriously considered to confirm the propeller is strong enough that it will not be broken

when it is rotating. The FOS graph can easily display the safety property. If FOS is not

less than 1.5, it means the propeller is strong enough. Figure 35 shows the FOS of the

propeller.

71
 ... fcl: ._, ' - t T-* ~ ~ MSYSII .O 'Miclow ... - IJ •

[) Iii ii. &. ®v. I Ill 3D· 'Ill e · [J · ill ~ ilo II R llo ll~ 1111 11 •" "' · 1 o· ~ ·
[II'] ..!. . ...!.. ...!.. ..:.!:.. ~ .! ..2.. .s aS. :! ,! ,.,
l=u ~,.:,..,. ,_,._ ~ ....... ..,_..; ....______

a..·-.---
-!
-
- .!. ....!... . J'_ .
-J! o!.

-·-- .......................
~oi..IIS.IC:,I,;;;,.L-...; ~...:::..~0111*1

. . . , .,...... (1)·)

u-

~
·-- ~=I
;::::--(4DW:......)
ilil;.lli!t.iiiliiiiiimmlllllll...._l.-.l
-

Figure 35: FOS ofthe 3-Blade Propeller

In Figure 35, the minimum FOS is 1.5 that displays at the left top comer of the FOS plot.

That means all of the values ofFOS are not less than 1.5, which means that this propeller

with the Titanium Alloy material is strong enough. There are two main methods to

change the strength of propeller design: One is to changing the material of propeller

designs; another method is change the geometry of the propeller design. Studying from

Figure 34, the connections of hub and propeller blades have the highest stress and can be

broken with high probability. A good hub design can reduce the stress at the connections,

such as adding much more material at the connections. Fillets can be added at the base of

blades to increase the strength. Figure 36 dispalys a fillet at one root of a propeller blade.

Figure 37 is the stress distribution that only adds fillets for this 3-blade propeller.

72
 Figure 36: Fillets at the Root of a Propeller Blade

~ 11 t li a ·'lll :~ • 5 IUl il • • "' " " I)"

- ~

"4i "
.
---
...~ ...!. ..!..
.......--....
....._. CUI

..._,_._,
....,._
~
Qa
..! ~
...._..,.,._
.sa aS. !
--
! .!.
-
..,!!... ~

!-......
' c!"~lj)

.,_
=-
._...,
~­.,_,
«-· .....,
·--
;::--(_..
-
. .-
w sa us

Figure 37: Stress Distribution with Fillets of the 3-Biade propeller

73
Figure 37 uses the same colore setting as the Figure 34. From Figure 37, the stress on the

propelelr is decreased with adding fillets. This result can be easily displayed with FOS

distribution, which is shown as Figure 38. From the left top of the FOS plot, the minimum

FOS value with fillets is 5.1 . The minimum FOS value without fillets is 1.5. That means

adding fillets at base of blades can increase the strength of a propeller design.

. ".
o w•• • ~ \91 1 11 :'11 ·111 • · · II • 'U II il lllo • • 1"1'11 " CJ " " "
..!.. · .!. · ......
--
...!. ..!..
cw.
~ ~
......
.! ~
at............. S cS I ! .! .! -...t1
-- -
.!. .!. ..! ...!.. . J!. .

·-..... ........................
01111-.------
- - - .... 1
--~_.CIWIIII

·-·--
1-
~

.,_........,
·-·d-•
(

;:::--(.......... )

..........
.,_
AI-
)
.. J5iliW&iiiiWI

Eduuttonal VtratOn Fot W'lwtrucbOnll Us• Only

Figure 38: FOS Distribution with Fillets of the 3-Blade Propeller

Figure 39 is the flow chart of the propeller simulation process.

74
 Open a propeller geometry in SolidWorks

Open the CosmosFloWorks tool and set up the simulation parameters

Run the simulation with all propeller blades and record the result

Run the simulation without propeller blade and record the result

Add these two results to get the final simulation result of propeller thrust

Use CosmosWorks to check the strength of the propeller design

Figure 39: Flow Chart of a Propeller Simulation Process

In Chapter 3, the method of generating a propeller foil section, a leading edge line and a

propeller blade is presented. Several blades are joined to a hub to create a full propeller

model. Based on the real propeller settings, a simulation model is created by the

simulation package CosmosFloWorks, which can simulate the rotation speed and the

motion of a propeller, and then to predict the thrust of the propeller. CosmosFloWorks

can also calculate the surface pressure of propeller blades, which can be used as the input

for FEA analysis application "CosmosWorks" to check the strength of the propeller

design.

75
Chapter 4

Case Study: Simulation of an AUV Propeller

As introduced in chapter 2, D'Epagnier designed an AUV propeller by OpenPVL code. ln

this chapter, the OpenPVL_ SW code was run with the same parameters as in

D 'Epagnier's study, and generated a propeller geometry file for SolidWorks. Because

OpenPVL and OpenPVL_SW have the same calculation methods of propeller design, the·

propeller geometry is the same. The main purpose of this case study is to verify that the

OpenPVL_ SW code is not only capable of generating a propeller geometry file for

SolidWorks, but also the CosmosFloWorks can achieve reasonable simulation results for

predicting propeller thrust. After running the propeller design function with the propeller

parameters in OpenPVL_SW, the OpenProp_Solidworks.txt file was created to generate

the propeller blade in SolidWorks. This file records the macro to automatically draw the

propeller blade geometry in Solid works. An abbreviated version of the macro is shown in

Appendix B. The propeller blade geometry is shown in Figure 40.

76
 '1!flo l<l liM t-<
D ..t flP~ ~
I. -b-
~ 11<
tloO
e Il!D·!'>l luJ·
..lf.I..!!J
..lf.I..!!J

~ N t):!j [l \ Gl
=.....-.---,---"---- -
Ill 0 Ql '(/ • 'II .
li-MP... ~ ... ~~'ncr ..,.,..., C\.r"WM

lJ
v
Q

).
- Figure 40: AUV Propeller Blade Geometry

After a propeller blade is created, a hub needs to be designe. Following the parameters of

the AUV Propeller, the diameter of the hub is 0.12192 m. In this case, a simple nose cone

with 0.06m height is provided to be the propeller hub. The nose cone is used to reduce the

inflow force when the propeller will be worked on underwater vehicles. The AUV

Propeller has three blades; therefore, another two blades will be added manually by the

SolidWorks function "Circular Pattern". The final propeller is shown in Figure 41.

77
 tJ
•;
'I

'll

11 'tdtrt'-'
1
T

Figure 41: Final AUV Propeller

After the propeller geometry is created in SolidWorks, CosmosFloworks is used for the

propeller simulation. In this simulation, the settings are as listed:

• Because the inflow wake velocity is too small compared with the ship velocity,

the simulation ignores the inflow velocity.

• The propeller is set to immovable, and the fluid is uniform and moves at 1 m/s

backwards to simulate the ship motion.

• The rotational speed of the propeller is 120 RPM.

• Gravity feature is considered, g=9.81m/s2 .

• Water is used in this simulation.

• The simulation tank is 1Omx 1Omx50m (widthxdepthxlength)

• The temperature of water is 293.2 K.

• The air pressure is 101325 Pa.

78
 • Assume the roughness is 0 micrometers.

• Pressure potential is considered.

• No cavitation is in the simulation

• Adiabatic wall is used.

The Figure 42 shows the simulated force result for the propeller with all three blades. The

simulation results are based on several calculation cycles. At the beginning, the

simulation has not enough calculation cycles, the results has large error, which shows as

the plot before 10 iterations. After several iterations, the results are much accurate. The

current value is the result after many cycles, so it well reflects the final simulation result.

The current value as shown in Figure 42 is 78.8 N.

[msolver: Default (l)(new002.5LDPRT) - (Goal plot 2] - 'A!~l ...J.QJ.lS
..:.l.ruJS

I Goal Name I Current Yalue_L AveJ..ged_YakreJ Proaress Delta Crterton

~ ~ GG v- c~ of F"'ce l -78.7606 N -78.3463 N ·104.962N -69.0J95 N 35.942ZN ON
I .!..1
Absolute Scal e(Auto Mln,Auto Max) lternlions
·69.0395 r--::;::;:;~-~===;;~=====;;~~---::~----;:-;;--------;:-;;--------:;-;;------;;-;;-----j
'l 0 20 30 40 50 60 70 80

1/

· 8549.37
l.!J_j
UJLiioai ~ --:---~~liiiiiiiiiiiiiiiiiiiiiiiiii-
Into
Ready Log_ :;41QfM· ,IterabOnS: 120

Figure 42: Predicted Result of Propeller Thrust With Blades

79
As introduced in chapter 3, another simulation, with no propeller blade, needs to be done

to consider the force effect from the water inflow on the propeller. Figure 43 displays the

simulation result with no propeller blade.

tmsolvet: Default (t)(ni!wOOV ilOPRT) ~ [Goal plot 2] ~':{ ~... ;; _

Goal Name C~.rrent VMle Avetac:~ed Value fltniml.rn Vak.Je Maxinum VO Proaess Deb Critenon COfl'VTlent

Absolute Scole(Auto Mln.Auto Max)

6.87782
6.6
6.4
6.2
6
5.8
5.6
5.4
5.2

4.8
4.6
4.4
4.2
4
3.8
3.6 \ I
V\./..,_,-; _ _ _ _ _ _ _ _ _ _ _ _ __ __ _ _ _ _ _ _ _ _
3.33471 L___:,c::::__;_ lterotions
_;,:::..::._~
I
10 20 30 40 50 60 70 80

~
O tnto I lil Log

Figure 43: Predicted Result of Propeller Thrust Without Blade

The current value for this no blade simulation is 3.9N. Thus, the propeller thrust from this

simulation is the sum of these two results, 82.7 N. The designed thrust ofthis propeller is

75N . There is only approximate 10% error between the designed and simulation result,

however; the simulation result is quite accurate for a prediction level.

Figure 44 displays the relation between RPM and thrust of the propeller at the vessel

speed 1m/s. The plot of RPM vs. thrust can be used for propeller design to avoid

cavitation and sympathetic vibration. In Figure 44, the change of thrust trends to be small

when the RPM is from large to low.

80
-------------- - -- - - - -- ----------

90
80
70
60
....
~so
'-
..c 40 1-+-seri es1 !
I-
30
20
10
0
0 20 40 60 80 100 120 140
RPM

Figure 44: Relation of RPM and thrust at vessel speed lm/s

In 2009, Dr. Vural finished several experiments to test several APC propellers and plot

the graph of RPM and thrust for each propeller. The testings are finished by an electric

motor, which is mounted on an RC engine that is mounted on a polycarbonate board.

Load cells are mounted on the polycarbonate board to test thrust [39] . The graph of RPM

and thrust from Dr. Vural 's experiments is shown in figure 45.

7 • P.PC 11 x5 .5 - - - - - - - - - - - - - - - - - - - - - - - ----------------------.
• P.PC 11 x7
6 ------ -------------------------------- -- -------------·o·------------------------
A APC 11 x8
0
5 L1. P.PC 12x6
• P.PC 12x8 o o
g 4 - o P.PC 13x6.5 --------- - - - ---------- - ------------------ -QJ ·---~ ------------------------------
c:B A
'lii A APC 13x8 o AA o • A
2 3 - --------------- -------------------& -- --- o______ A -1P----• -• ------ ____________ _
-{:_ <> APC 14x7 o ~ AA~ •• A A L\. • •

2 _ o APC 15x6 ----------------------------~--o5>..~---~9--.--~~-_..t:P...~--~-----------­

r:F o# ~o. ·~ • A •
---- - ------ - --- - --------- - --- - __ _c;P_<r -t:r),_,:~ . j.tl'!.-- ~---------------- -----------------
0 e~l ~·
O w---~--~~~~----~--~----~--~----~--~----~--~
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000
RPM

Figure 45: Plot of RPM and Thrust from Dr. Vural's Experiments [39]

81
The plots of RPM and thrust from our simulations and Dr. Vural's experiements have a

similar trend, which prove that our simulation results for the relation of RPM and thrust

are reasonable. The plot of RPM and thrust can calculate the RPM when the required

thrust at certain vessel speed is known. RPM is a main factor of cavitation, which will

destroy propeller and cause noise and vibration; low RPM is used to avoid cavitation.

RPM is also a factor of sympathetic vibration between propeller and vessel. Certain RPM

of propeller has its own frequency. If this frequency is equal to the natural frequency of

vessel, sympathetic vibration, which can destroy propeller and vessel, will be created. If

the thrust is known, this plot can provide the RPM information for designers to avoid

cavitation and sympathetic vibration.

Another two propellers were developed using OpenPVL_ SW with the designed thrust of

55N and 95N (other parameters are the same as the AUV propeller). Simulations were

used for these two propellers to confirm CosmosFloWorks can provide a good predicted

thrust result for propellers. Table 6 displays the thrust simulation results of the two

propellers.

82
 Table 6: Simulation Results of the Three Propellers

Simulation

Designed Simulated Domain Size

Error %
Thrust Thrust (widthxdepthx

length)

Propeller 1 55N 56.3N 2.38% 8mx8mx30m

Propeller 2 75N 82.7 N 10% 10mx10mx50m

Propeller 3 95N 99.5N 4.77% 15mx15mx70m

After the CFD simulation, CosmosWorks is used to check the strength of the propeller.

Figure 46 displays the surface dynamic pressure on the AUV propeller blades with 75N-

designed thrust. The dynamic pressure is the source to provide the thrust for a propeller

design. Each blade provides the same force, that's why the distribution is symmetrical.

83
 Figure 46: Surface Dynamic Pressure Distribution of the AUV Propeller

As stated in Chapter 3, Cosmos Works uses the surface pressure results as an input to

check the strength of the propeller design. In this case, the Titanium Alloy was used as

the propeller's material with yield strength 1.03x109 N/m2 . Other materials can also be

used in the propeller design. The Titanium Alloy was only used as an example. The rule

to choose a material is to make sure the propeller design has strong enough strength, and

to be cheap as possible. Figure 47 displays the stress results of the propeller.

84
 ~ it II • "'ci1 1111 11 • 1"1'1 • o· ~ ·
~ ~ ! ! ~
--- -
! ~ ~ ~ >I
~·a.....·
?I

Educattonat Vertlon. For lnltructtonaJ Uu Ontv

Figure 47: Stress Distribution of the AUV Propeller

In Figure 47, the areas near the hub have the higher stress, and these areas are needed to

be seriously considered to make sure the propeller is strong enough. Stress on the

propeller will ultimately cause deflections, which is also important when considering a

propeller's acceptability. High deflection on a propeller will change the flow around the

propeller and loads on it. Low deflection is desirable for a propeller design.

Cosmos Works can output the displacements related to the stresses generated by the

fluid/blade interaction. Deflections were briefly reviewed and seen to be minimal but the

simulation values for stress and deflection should be studied in more detail in a further

study. The analytical results should be validated through experimental work, which has

been recommended in the conclusions of this thesis. Figure 48 displays the deflection of

the AUV propeller.

85
 8 N 1\ f/l ill llllo • • ,." " O " 'C "

-- -
J.!!. o!. .! ...!.. . J!. .

Edue.tlonaiVersion Fot lnstruct.onat Uu Onli

Figure 48: Deflection of the AUV Propeller

From figure 48, the highest deflection is at the top of the propeller blades, and the value is

0.0042 m. This deflection value is acceptable for a high propeller strength property.

Figure 49 shows the FOS of the propeller.

86
 () IiJ.. ___
I .!. . ..!.. ..!.. ..:.. ~
v. I
.! ~
ll!io·,_ e · IO · ~

.s
lit II Ill
o.S. :t .! .! .!. .... .!. a!. .!.
• o· " "
...... o.t..............
Q.t ........ ~..--

-- ...,.... ~-.., 1=---..=.:-,...a..t

. ..., . ( ...... (1)-)

,..,....................
~.- ...~- .
~=--
·-·
·--
cfl .... - ·1

...............
r:1
liJ
-~ ,
llll-
161;-""
wl!Eif.iiilliiiiBmmllll
.....m.,

'. stan • "

.
•
.
Figure 49: FOS of the AUV Propeller

In Figure 49, all ofthe values ofFOS are bigger than 1.5, which means that this AUV

propeller with the Titanium Alloy material is strong enough.

Chapter 4 outlines the way to design an AUV Propeller using OpenPVL_SW and

SolidWorks. The exercise demonstrates that SolidWorks is not only able to support the

design of a desired propeller, but also that CosmosFloworks can be used to predict the

thrust result of a propeller and generate a pressure profile. CosmosWorks can then use the

pressure profile to analysis propeller strength by FEA to ensure the propeller design is

strong enough when it works. If the strength is weak, redesigning the propeller geometry

or changing other higher strength material can be used to increase the propeller strength.

87
Chapter 5

Case Study: Simulation of a Marine Propeller

Provided by Oceanic Consulting Corporation

Oceanic Consulting Corporation is a company, which provides marine related services

about design evaluation and testing. ln this case study, Oceanic Consulting Corporation

provided a propeller geometry and tested thrust result for our simulation validation using

CosmosFloWorks. ln a propeller testing, the results may be affected by many factors.

CosmosFloworks cannot simulate all of these factors. That is the main reason why there

arc some differences between the testing results and simulation resu lts. The main purpose

of thi s case study is to verify CosmosFloworks can provide a believable result of

propeller thrust. 1\ known propeller geometry was provided by Oceanic on ulting

Corporation, as one of doing the simulation validation. The company had already

determined the experimental values ofthe propell er thrust. The validation would come by

seeing ifthe simulation value is close to the experimental one. This would mean that

Cosmosrloworks is a feasible tool for predicting propeller thrust.

88
-------------------

The propellct: has four blades, 0.15m diameter, rotates at 900 RPM and was tested in 0.5

m depth tank. The ship velocity is 1 m/s. The propeller geometry is provided by Oceanic

Consulting Corporation with a SolidWorks file, which is shown in Figure 50. r<igurc 51 is

the side view of the propeller geometry used in this simulation.

- a"•"'

Leading Edge / ose Cone

Figure 50: Propeller Gcomctry-4 Blades

Pressure Face

r<igurc 51: Sideview ofthe Propcllcr-4 Blades

89
In this simulation, two nose cones are added to the propeller in order to reduce the inflow

force on the propeller. As introduced in chapter 3, the inflow force on the propeller is the

main enor between the simulation and experimental testing results. Thus, deceasing the

inflow force is a way to increase the accuracy of the simulation results. In order to fast

simulation, designers can use two nose cones simulation method to predict propeller

thrust without using the no-blade method that is introduced in Chapter 3. It is only a

method for rapid simulation, however; the simulation method as introduced in Chapter 3

is preferred. The rotation direction of the propeller is shown in Figure 51. The propeller is

et to be immovable, and the fluid moves at I m/s in the direction shown in Figure 51 to

simulate the ship motion. When the propeller is rotating, water will create force on

pressure faces, and provide thrust for ship movement.

The settings of this simulation arc listed as below:

• The simulation tank is 1mx 1mx I Om (widthxdepthxlength)

• The propeller is immovable, and the fluid moves at 1 m/s backwards.

• The propeller is operated at 900 rpm.

• The depth of propeller center in water is 0.5 m.

• The temperature of water is 293.2 K .

• The air pressure is 101325 Pa.

• Gravity feature is considered, g=9.8 1m/s 2

• The roughness of the propeller surface is 0.3 micrometer.

• Pressure potential is considered.

90
 • Assume no cavitation is in the simulation

• Assume adiabatic wall is used.

Figure 52 displays the result of the propeller thrust with all blades. The fo rce is 29.0

mpellerOS.SlDASM) · (Goo! plot I] ' •'\:,"!Oll(

I ,.J file C~lculabon ~leW tnsetl ~ ~

• " > c ' * ~ fo ~ ··~~ • 1' '

Goal Name ClKrri Value Averaoed Value Minimum Value Maxlnum Value PrOQI'ess Delta Critenon Comment
. GG X·, .. ·29.0361 N ·30.3871 N ·32.6617 N ·27.8<22 N I Achieved 5.01951 N 8.66473 N

Ah so1utr. Sct~ le[Autn Mln,Auto Max) Iterations

-16.0117 I ~-
10 20 30 40 50 60 70 on

(
-685 1.88

.!.L.J
0 lnlo ]I_;J_Gooijllot 1

Figure 52 : Predicted Propeller Thrust - 4 Blades

As introduced in chapter 3, another simulation needs to be done to consider the force

effect of the water inflow on the propeller. Figure 53 displays the propeller geometry with

no blades. When water is flowing, forces are created on the nose cone 1. Because the

geometry of nose cone 1 is symmetrical, the summation force of these multidirectional

forces is the opposite direction of the thrust of the propeller.

91
 Inflow
Force

Nose Cone Nose Cone 1

.. .•
--~~~~~--=~~~----~--~--~--~--~--~~------~~~----~

Figure 53: Propeller Geometry without blade

In this simulation without blades, all the parameter settings are the same as the simulation

with blades. Figure 54 shows the simulation results without blade. The force is 0.2

92
lljsolver: Default (l)(propeller05.5LDASM) - [Goal plot Z] ~

Delta Crterlon Comment

0.00 13478 1 N 0 .001349 N Checl<klQ criteria

Ab solute Scale(Auto Min.Autn Max)

O.JO:l11 2

0.2
, ........____
0. 100 277 • -
__________l!!__ ----~----- ------1Q
_::-:::-:::::--=,-,::::::.=~-~====::::::::==========-
40 50 60
===--
7 11
------------------- -~
- - -·uo -
O lrlo Lf11 Log Jl /"'_ Goal ~~ 2_
Roady Solver 1s hntshed [Iterations : 133

Figure 54: Predicted Propeller Thrust for the Propeller without Blade

The force from the simulation with all blades is 29.0 , and the force from the simulation

with no blade is 0.2 . Thus, the propeller thrust from this simulation is the sum of these

two results, 29.2 . The real testing result of this propeller from Oceani c Consulting

Corporation is 34.9N. There is approximate 16% error between the real and simulation

results. The result has 16% error, however, it is quite accurate for a prediction level.

As introduced in Chapter 3, CosmosFloWorks can provide the surface press ure

distribution to be as an input of FEA to check the strength of a propeller design. fi gure 55

displays the dynamic pressure distribution of the propeller. The dynamic pressure is the

source to provide the thrust for a propeller design. Each blade provides the same force,

that's why the distribution is symmetrical.

93
 ~Q C+ QJ · I@CJJ CJJ .
,:~==~=-~~~~~~~~~~
& o ~ Q c • · ~ ~ ~-~~~
~ m ______ I •

Figure 55: Surface Dynamic Pressure ofthe Propeller-4 Blades

Figure 56 displays the stress distribution results in the FEA analysis.

. ~.

D 101 11 111! !1> '9 · ['I "' . lii !D ·!ll • · 11:1 · 'll< ioi> I } P IJ G Q <?> ~:> II e • III '*' II:I .C I
s ,..
-··
:II' llll ll ..
....1,. . .!, - ~
! ... """"'-
'--«.~,~
RlltD '-" ~
~
"'*-"•·..
o.t:.ctbl
I'Mbnl • ..u.tlarl •

EducatlonaJ Version For Instructional u .. Only

Figure 56: Stress Distribution of the Propeller-4 Blades

94
In the Figure 56, it can be seen that the higher stresses occur at the tips of the blades. In

this propeller design, the hub has an elliptical geometry to reduce the flow force and

much more material is applied between the hub and blades can provide higher strength

potential. Figure 57 displays the deflection of the 4-blade propeller.

~ ,.. l.l !B· JII .. · [] · :1- ..

-==-
!J'
•"' ... -
...!...

- ..
,_
. . . . , ........ (1 ..)

~­
----m_..
c - ...

~
II ·
~

I :IDI
·--
d"-'--1
~ ~I
1111 ·
~~(-QIW ......)

··---1(. . . . .)
~-
c~Chldll(.f'OS-)
. liiNiliiidiiiiGiiilbiJ

Educabonal Vtrs•on For lnstnJetion.t Ute Onty

Figure 57: Deflection of the Propeller-4 Blades

In figure 57, the highest deflection is at the tip of each propeller blade, and the value is

0.000046 m. This deflection value is small enough that it is not a concern from a

functional viewpoint.

95
FOS results are plotted to make sure this propeller design is strong enough. In this FEA

analysis, Titanium Alloy was used as an example to be as the propeller's material. Figure

58 displays the FOS results of the propeller.

· · ~' 'I d • lll '~ P D $ 0

•No '
,... .,£1 lY
............. " ~

-..... -
~ ... ~c-...c,....... ~

" ~
C !'*'-
t - · ........
t"...,,, (l).)

\tiMiflll'l_...

·-- ot- ·...........

d ...... I

..-
~,

Jl,-

. ,,.,..,«............_)
IE' J!E'!iii!hiii113

Figure 58: FOS of the Propeller-4 Blades

In this FOS results, all the values are bigger than 1.5, which means the propeller is strong

enough with the Titanium Alloy material when it is rotating.

In this simulation study, the propeller geometry was provided by Oceanic Consulting

Corporation. However, engineers often need to design based on a mission profile, and

then to do a simulation. OpenPVL_SW can be used to design a propeller, create the

propeller geometry in SolidWorks, simulate the propeller thrust in CosmosFloWorks, and

check the strength of the propeller by Cosmos Works. Figure 59 is the flow chart of the

96
procedures of a propeller design and simulation using OpenPVL_ SW, SolidWorks,

CosmosFioWorks and CosmosWorks.

Design a propeller blade using OpenPVL_SW:

• Parameter analysis
• Generate foil sections points using OpenPVL_SW
• Generate a propeller blade using the OpenPVL_ Solidworks.txt file

Design a hub and add other blades to complete the propeller geometry in
SolidWorks

CFD simulation:
• Open the CosmosFloWorks and set up the simulation parameters
• Run the simulation with all propeller blades and record the result
• Run the simulation without propeller blade and record the result
• Add these two results to get the final simulation result of propeller
thrust
If the thrust is desired as designed, then go to the FEA for the strength
analysis; if the thrust is not the desired as designed, then go back to
redesign the propeller parameter or the hub.

FEA simulation:
• Open the CosmosWorks and create a static study
• Import the pressure result from CosmosFloWorks into the static
study
• Select propeller material
• Set bonded and centrifugal features for the propeller
• Check the strength of the propeller by the stress distribution and
FOS results
If the propeller has high strength, then be ready for fabrication to test; if the
propeller has low strength, then go to redesign the hub or use higher
strength property material for the propeller.

Figure 59: Flow Chart of the Procedures of a Propeller Design and Simulation using

OpenPVL_SW, SolidWorks, CosmosFioWorks and CosmosWorks

97
Chapter 6

Propeller Fabrication by Rapid Prototyping

Several design and simulation iterations may be required before the designer is satisfied

that the propeller geometry is appropriate. Once satisfied, the designer will need to

fabticatc a working prototype for testing. Rapid prototyping (RP) is the recommended

approval to minimize the time and cost to arrive at a suitable prototype. As introduced in

chapter 2, there arc six rapid prototyping teclmologies arc introduced. All of these

technologies can be used to produce a propeller. The main differences between RP each

technology arc the materials to manufacture parts and the machine accuacy. Selection of a

suitable RP technology for a propeller depends on the accuracy and the strength required

for the specific working environment. Stratasys Fusion Deposition Modeling (FDM) 2000

is used in this thesis for propeller fabrication due to its accessibility at Memorial

University. /\s stated the introduction in chapter 2, FDM technology is fast, at low cost

and quite accurate for a testing level; however, the materials that can be used in PDM arc

limited with ABS, PC and wax. These materials have poor strength property, which limits

the strength ofFDM prototypes.

98
The Stratasys I~ OM 2000 RP process requires the input of an .STL file, which is a fonnat

used by Stcrcolithography software to generate information need to produce 30 models

on Stcrcolithography machines. Specifications of the roM 2000 used in this study arc

listed in Table 7, and the FOM 2000 is shown in Figure 60.

Table 7: Stratasys roM 2000 Specifications l40]

Build Size 10 x !Ox10 in

Achievable Accuracy +/- 0.005 in

Acrylonitrile Butadiene Styrene (ABS)·

Medical grade ABS; Methyl methacrylate

Modeling Materials A.BS ; Polycarbonate plastic; lnvestmcnt

casting wax

Layer Width Ranges from 0.0 I 0 to 0.100 in

Layer Thickness Ranges from 0.002 to 0.030 in

.STL files arc processed using Quicks! icc

Software
Version 6.0

99
 hgure 60: Stratasys FDM 2000

After a propeller solid model is created in SolidWorks, the propeller geometry can be

exported as an .STL file. When an .STL file is input into Quickslice Version 6.0, a model

structure is automatically created for the file, shown in Figure 61.

100
 Figure 61: Screenshot of Propeller Blade Generated by Quickslicc Software

The FDM printable file can generate both a total volume calculation and a completion

time estimate before the production. The following lists are the steps for the fabrication

using Stratasys FDM 2000.

• Start Quickslice and open the .STL file.

• Choose tips of material extrusion nozzle. Small tip can provide higher strength

and smooth surface finish, however; it will take much more time and material.

Because the blades of a propeller are thin and need to support high stress, small tip

is preferred for propeller fabrication.

• Set up the orientation of a model in Quickslicc. The two options of moving and

rotating can be used for the orientation and make sure the model is in the

101
 construction area, which is lined in the software. Because the FDM builds up

models in layers along the z-axis, the orientation will affect how the final model

comes out. The ideal orientation is stable, minimize support material, and place

any intricate structures facing up so that support material cannot mar them . For

propeller models, it is better to place the hub structure facing up, as shown in

Figure 61.

• Create slices to build up models using Quicks] ice. The thickness of slice is from

0.002 in to 0.030 in. Low thickness can provide strong structure and good surface

finish, however; it will take much more material and time.

• Create support that will allow parts to be built in it.

• Create tool path for the fabrication. The toolpath of parallel straight lines are

usually used in fabrication. The way of several curved parallel lines for the

propeller blades' edges and straight lines for the middle area is suggested for

propeller blades fabrication.

• Create base and then save as an .SML file, which is a solid model file that wi ll be

sent to the FDM machine

• Turn on the StrataSys FDM 2000 and warm up the machine. The required

temperature for ABS is 270° (model material temperature) and 265° (suppOii

material temperature).

• Check FDM material loading to verify that the model and suppOii materials arc

loaded correctly.

• Send the .SML file to Stratasys FDM 2000 to produce the part.

102
Material options for Stratasys Fusion Deposition Modeling (FDM) 2000 include

Acrylonitrile Butadiene Styrene (ABS); Medical grade ABS; Methyl methacrylate ABS ;

Polycarbonate plastic (PC); Investment casting wax. Wax is too weak to produce a

propeller. ABS can be used to produce a propeller, and can be easily filed down to a

smooth finish using sand paper, if necessary. PC is more rigid than ABS, but it is

expensive. Medical grade ABS and Methyl methacrylate ABS are the ABS/PC blend

material to reduce the cost if extreme strength is not necessary, and can achieve specific

purposes. Methyl methacrylate ABS is often used as transparent glass substitute. Medical

grade ABS can be used for surgical instruments, diagnostic devices and drug deli very

systems 1341. The material will be selected based on the cost and ri gidity requirements.
ABS is economical but with low rigidity. PC is the most expensive with very high rigidity.

Methyl methacrylate ABS and Medical grade ABS have moderate rigid and cheaper than

PC. As introduced, PC is the best material to fabricate a propeller with hi gher strength .

Figure 62 is an example ofFDM propeller prototype with the material of ABS.

Figure 62: A Propeller Fabricated by FDM

103
Due to the accuracy limitation of the FDM machine, the edge and the surface of the

propeller model are not very smooth. Glue can be used to cover the propeller to achieve a

smooth propeller body. Epoxy is a good choice for a propeller surface coating, because

epoxy can provide a thin wall around the surface and increase the strength of material s

r41]. [n order to avoid unequal coating thickness, it is better to coat one blade at a time
and keep the blade flat when the glue is drying. When the epoxy is fully dried, sand paper

can be used to make the blades much more smooth. As introduced before, PC is the

strongest material to produce a strong propeller using FDM. PC coating with epoxy can

provide higher strength than PC only. The detailed data for the strength of the PC coating

with epoxy material will need some experiments in the future work. If the material is still

not strong enough for some specific cases, the FDM propeller model can be used as a

casting mould to produce a metal propeller.

In chapter 6, Stratasys Fusion Deposition Modeling (FDM) 2000 is introduced. The

method of producing a propeller using Stratasys 2000 is also presented. The material that

can be used in Stratasys 2000 is listed, and PC is the best material for a propeller

fabrication using FDM. Coating with glue can provide smooth and higher strength

property for a propeller model.

104
Chapter 7

Conclusions and Future Work

7.1 Conclusions

The described rapid marine propeller design process was developed to facilitate a

propeller design from simulation through to prototyping and testing processes. The

OpcnPVL_ SW code was developed to generate the propeller geometry for automatic

input into the CAD software, SolidWorks. SolidWorks includes the CFD simulation

capability for propellers using the CosmosFloWorks package. This study has proven that

CosmosFloworks can generate a reasonable prediction for propeller thrust. CosmosWorks

can then be used to check the strength of a propeller design using FEA analysis.

The original OpenPVL code was able to generate the propeller design geometry for the

CAD software, RHINO. This geometry is generated as the sets of points that arc located

on the foil sections of the propeller blade. Users then had to usc several commands to plot

out the geometry of the propeller in RHINO. The OpenPVL_ SW code extended this

capacity to generate the propeller geometry without any manual commands in

105
SolidWorks. This modification will clearly save users' time, especially when a number of

design iterations need to be implemented. Solidworks allows users to desig11 a specific

propeller hub and create a 3D-printable file for rapid prototyping to produce a propeller

for a model testing.

The CosmosFloworks package allows users to simulate the propeller working in a

specific fluid envirorunent through CFD. The simulation helps the designers to predict the

thrust of the propeller as part of the design cycle. The CosmosWorks package will usc the

pressure results as an input ofFEA to check the propeller strength.

7.2 Future Work

• Add propeller hub design function in the OpenPVL_ SW code

• Coating with glue can increase the strength of ABS material. Future experiments

can be conducted to find out the detailed data for the increased strength.

• Usc SLS rapid prototyping technology to fabricate an ideal marine propeller and

test its pcrfonnance

• Test propeller's thrust and compared with the CFD result

• Test stress at the root of the propeller blades

• rabricate a propeller with two methods. One is produced by CNC machining with

cutting off material to left a solid part. Another one is produced by melting

material layer to layer to fabricate the propeller part. Test the stress of the two

propellers with different fabrication methods.

• Much more deflection analysis will be applied in the future work.

106
. - - - : - - - - -- - - - - - - - - - - - - - - - - - - - -

References

[1] Burtner, E. , 1953, "A Relationship for Preliminary Propeller Diameter", Journal of

the American Society ofNaval Engineers Inc. , Vol. 65, o. 3.

[2] Betz, A. , 1910, "Schrauben Propeller mit Geringstem Energie Verlust", translated

from German by T. Hofmann and N. Bose, " Screw Propellers with Minimum

Energy Loss", 1990, Ocean Engineering Research Center, Memorial University of

ewfoundland, Rep. No. OERC90-HYD-LR-02., pp. 20.

f3 -J Goldstein, S. , 1929, "On The Vmiex theory of Propellers", Proceedings of the

Royal Society, Series A, Vol. 63 , pp.440-465 .

[4] Glaueii,H., 1947, "The Elements of Airfoil and Airscrew Theory", Cambridge

University.

[5] Lerbs, H.W., 1952, "Moderately Loaded Propellers with a Finite Number of

Blades and an Arbitrary Distribution of Circulation", Society of Naval Architects

and Marine Engineers, Vol.60, pp. 73-123 .

f6l Kerwin, J .E., 200 I , "Hydrofoils and Propellers", Massachusetts Institute of

Technology.

f7] Eckhardt, M.K., Morgan W.B. , 1955, "A propeller Design Method", Society of

aval Architects and marine Engineering, Vol. 63 , pp.325-374.

f8] Erickson, L. L., 1990, "Panel Methods-An Introduction". ACA Technical Paper

2995, California

107
[9] IIess, J .L. and Valarezo, W.O. , 1985. "Calculation of steady flow about propellers by

means of surface panel method", Research and Technology Douglas Aircraft.

Company, Long Beach

II OJ Ramachandran, P., Rajan, S. C. and Ramakrishna, M., 2003, " A Fast, Two-

Dimensional Panel Method". Society for Industrial and Applied Mathematics

[11] D'Epagnier, P .K., Chung, H., Stanway, J.M., Kimball, W.R., 2007, "J\n Open

Source Parametric Propeller Design Tool", Cambridge.

!"12] Liu, P., Islam, M.F. , Searle, S. , 2006, "Design and Optimazation of An Icc Clas

propeller Under Shallow Water, Semitunnel Hull and Heavy Load Conditions",

111
Proceeding of OMAE2006, 25 International Conference on Offshore Mechanics

and Arctic Engineering, Hamburg, Gennany.

p 3] Muckle, W ., 1941, "Stresses in Propeller Blades", The Shipbuilder and marine

Engine Builder, Vol. 47, o. 388, pp.336-341.

[14] Morgan, W.B. , 1954, "An Apimate Method of Obtaining Stresses in a Propeller

Blade", David Taylor Model proxBasin, Report No. 919, Washington D .C. , pp.18.

rt5l Ghose, J.P., Gokarn, R. P., 2004, "Basic Ship propulsion", New Delhi.

[16] Hofmann, T., 1991 , "Marine Propeller Design, Tests and Manufacture",

Memorial University.

[1 Ti Kerwin, J.E., 2001, "Hydrofoils and Propellers". Cambridge, MIT CopyTech.

r 18.1 Chung, H. , 2007, "An Enhanced Propeller Design Program Based on Propeller

Vortex Lattice Lifting Line Theory", Master Thesis, Massachusetts Institute of

Technology.

119.1 Marzocca, P., 2009, "The ACJ\ Airfoil Series". Clarkson University

108
[20] D'Epagnier, P.K., 2005, "A Computational Tool for the Rapid Design and

Prototyping of Propellers for Underwater Vehicles". Master Thesi , Massachusetts

Institute ofTechnology

p 11 Farin Gerald E., 1993, "Curves and surfaces for computer aided geometric

design: a practical guide", Boston, Toronto

122] "CAD soft wares history', www. CADAZZ.com, Accessed 10 Jan201 0.

123] Lombard, M., 2009, "SolidWorks 2009 bible", Indiana,

l24j Anderson, J.D., 2009, "Computational fluid dynamics: an introduction". Berlin

125] Chaskalovic J ., 2008, "Finite element methods for engineering sciences". cw York

[26] Versteeg, H. K., 2007 "An introduction to computational fluid dynamics: the finite

volume method". Toronto

p7] Dymarski, P., 2008, "Computations of the propeller open water characteristics using

the OLAGA computer program. Predictions of the cavitation phenomenon". Ship

Design and Re carch Centre S.A.

128] icholson, D. W., 2008, "Finite Element Analysis: Thetmomechanics of Solids".

Boca Raton

[291 Richard L. Van Horn, 1971, "Validation of Simulation Results". amegie Mellon

Uni ersity

130] E. Paul DeGarmo, J. Temple Black Ronald A. Kohser, 1988, "Material and

processes in manufacturing". ew York

131 J Flinn R.A. , 1963, "Fundementals of Metal Casting", Addison Wesley Pub. Co.

Inc.

109
[32] Ali Kamrani, Emad Abouel Nasr, 2006, "Rapid prototyping: theory and practice".

cw York

[33-1 D.T. Pharo, S.S. Dimov, 2001, "Rapid Manufacturing: the technologies and

applications of rapid proto typing and rapid tooling". New York

[34] Hague, R.J.M. , Reeves, P.E., 2000, "Rapid Prototyping, Tooling and

Manufacturing". United Kingdom

[351 adooshan, A.A., Daneshmand, S. and Aghanajafi, C., 2007, "Application of RP

Technology with Polycarbonatc Material for Wind Tunnel Model Fabrication".

World Academy of Science

[36-1"Metal Lascr-Sintering."

http://www.cos.info/en/homc.hbnl. Accessed 28April201 0.

r3TI Karim, M. M., Suzuki, K. and Kai , H., 2004, "Optimal Design ofllydrofoil and

Marine Propeller Using Micro-Genetic Algorithm". Journal of aval Architecture

and Marine Engineering

f38] Lee, S., 2008, "Ice Controllable pitch propeller-Strength Check based on lACS Polar

Class Rule", Texas, USA

r39] Vural , M. , 2009, "Experiment results ofRPM and Thrust for Various APC

Propellers". Experiment Report, Illinois Institute of Technology

l40j "StrataSys FDM 2000 Specifications."

http://maclabs.ucsd.edu/alex/rapid prototyping.html. Accessed 1OAUG2009.

r411 "Devcon Epoxy."

http://www.devcon.com/. Accessed 30April201 0.

110
Appendix A: The Part ofOpenPVL_ SW MATLAB Source Code for Generating

Propeller Geometry in SolidWorks

111
The original OpcnPVL code is available at

http://web.mit.edu/openprop/www/Download.html. The idea of this OpenPVL SW code

is use Solidworks to instead ofRihno with the same calculation methods. Based on this,

replace all of the "rihno" words with "Solidworks" in the original code, which includes:

• Make_rihno_flag=1 changes to Makc_Siidworks_ flag= l in Matlab code line

727

• Geometry( ... Makc_rihno_ flag ... ) changes to

Geometry( . .. Make_Solidworks_flag ... ) in Matlab code line 730

• Function [J=Geomctry ( ... Makc_rihno_flag . .. ) changes to Function []-=Geometry

( .. . Make- Solidworks- flag . ..) in Matlab code line 1217

The part of Matlab code to generate a propeller blade in Solidworks is integrated in the

Matlab code line 1514 to instead of Rihno code. The logic flowchart of this solid works

code displays as below:

112
 Create all of the function commands at the beginning of the solidworks
macro

Use the calculation results in the original OpenPVL and sketch the first foil
section, which is located at the root of the blade. (get points first and then b-
spline connect all of the points)

Sketch all the other foil sections in the order of from root to tip of the
propeller blade.

Connect all the points, which are located in the leading edge line to be as a
reference line for lofting future

Selecting all the foil sections and the reference line to loft them to generate a
propeller blade

Hide all the points and curves for nice look

113
The Solidworks code is shown as below: Comments are following after the symbol %.

% -------------------------------------------------------- Make So lidworks files

if Make_ Solidworks_flag
%------Make _Solidworks.txt, with coordinates for the entire propeller
filename_ Solidworks = strcat(filename,'_ Solidworks.txt'); %create the soildworks
macro file
fid = fopen(filename_Solidworks,'w'); % open the file for writing
fprintf(fid,'Sub mainQ\n'); %Macro starting command
Start macro and declare objects
for the macro

fprintf(fid,'Dim swApp As SldWorks.SldWorks\n'); %declare swap is a

solidworks object
fprintf(fid,'Dim swModel As SldWorks.Mode1Doc2\n'); % declare
swModel is a document object
fprintf( fid,'Dim swSketchPtO As SldWorks.SketchPoint\n'); % declare sw
SketchPtO is a sketch point object
fprintf(fid,'Dirn swGuidePt(21) As SldWorks.SketchPoint\n'); % declare
swGuidePt(21) is a sketch point object
fprintf(fid,'Dim swSketch(21) As SldWorks.Feature\n'); % declare
swSketch(21) is a feature object
fprintf(fid,'Dim swFeatCurve(21) As SldWorks.Feature\n'); % declare
swFeatCurve(21) is a feature object
fprintf( fid,'Dim swGuideCurve As SldWorks.Feature\n'); % declare
wGuideCurve is a feature object
fprintf(fid,'Dim z As Long\n'); % declare z is a 32-bit number
fprintf(fid,'Dim bRet As Boolean\n'); % declare bRet is a ture or false
value

fprintf(fid,'Set swApp = CreateObject("SldWorks.Application")\n'); %open Solidworks

sesswn
fprintf(fid,'Set swModel = swApp.ActiveDoc\n'); %grab the current document
fprintf(fid,'swModel.SetAddToDB True\n'); % set to ture allows data to be directly
added to Solidworks model

fprintf(fid,'swModel.ClearSelection\n'); % set to clear anything currently selected

Start 3D sketch to generate
~------ propeller foil sections
~

fprintf(fid,'"start 3D sketch\n');
fprintf(fid,'swModel.Insert3DSketch\n'); % insert a new 3D sketch for the points
fprintf(fid,'ReDim swSketchPt(40) \n'); % define 40 points in the sketch

114
 Generate points on propeller
foi l sections and connect points
to generate foil sections

for j= l :2*Np %for each point for a foil section

fprintf(fid,'%s %g %s (%f, %f, %f)\n','Set swSketchPt(', j , ') =
swModel.CreatePoint2', X3D(lj, l), 0.9*Y3D(lj, l), 0.9*Z3D(l j, l)); % print poits data

end
fprintf(fid,'Set swGuidePt(O) = swSketchPt(l) \n'); % set the first reference line point

fprintf(fid,"'exit 3D sketch \n');

fprintf(fid,'swModel.lnsert3DSketch \n'); % insert the sketch in Feature object

fprintf(fid,'Set swSketch(O) = swModel.FeatureByPositionReverse(O) \n'); % define

the reference line point in the feature object

fprintf(fid,'swModel.ClearSelection \n'); % set to clear the current selections

fprintf(fid,'For z = 22 To 39 \n'); % define the points 22 to 39

fprintf(fid,'bRet = swSketchPt(z).Select2(True, 0) \n');
fprintf(fid,'Next z \n');

fprintf(fid,'For z = 1 To 20 \n'); % define the points 1 to 20

fprintf(fid,'bRet = swSketchPt(z).Select2(True, 0) \n');
fprintf(fid,'Next z \n');

fprintf(fid,'swModel.lnsert3DSplineCurve (True) \n'); % use spline to connect the

defined points

fprintf(fid,'Set swFeatCurve(O) = swModel.FeatureByPositionReverse(O) \n'); %

define the first refernce line point position in Feature object

fprintf(fid,'swModel.ClearSelection \n'); %clear anything currently selected

for i = 1:Mp % for each section along the span

fprintf(fid,"'start 3D sketch\n');
fprintf(fid,'swModel.lnsert3DSketch\n');
fprintf(fid,'ReDim swSketchPt(40) \n');

for j = 1:2*Np %for each point for a foil section, the each line has the same
function as introduced as before
fprintf(fid,'%s %g %s (%f, %f, %f)\n','Set swSketchPt(',j, ') =
swModel.CreatePoint2', X3D(ij ,l), Y3D(i j, l), Z3D(ij,l ));

115
 end
fprintf(fid,'%s %g %s \n', 'Set swGuidePt(', i, ') = swSketchPt(l)');

fprintf(fid,"'exit 3D sketch \n');

fprintf(fid,'swModel.lnsert3DSketch \n');

fprintf(fid,'%s %g %s \n', 'Set swSketch(', i, ') =

swModel.FeatureByPositionReverse(O)');

fprintf(fid,'swModel.ClearSelection \n');

fprintf(fid,'For z = 22 To 39 \n');
fprintf(fid,'bRet = swSketchPt(z).Select2(True, 0) \n');
fprintf(fid,'Next z \n');

fprintf(fid,'For z = 1 To 20 \n');
fprintf(fid,'bRet = swSketchPt(z).Select2(True, 0) \n');
fprintf(fid,'Next z \n');

fprintf(fid,'swModel.Insert30SplineCurve (True) \n');

fprintf(fid,'%s %g %s \n', 'Set swFeatCurve(', i, ') =

swModel.FeatureByPositionReverse(O)');
Generate leading edge line as
fprintf(fid,'swModel.ClearSelection \n'); the reference line for loft
end function

fprintf(fid,'"create guide curve \n'); %create reference line

fprintf(fid,'For z = 0 To 20 \n'); % define the point 0 to 20

fprintf(fid,'bRet = swGuidePt(z).Select2(True, 0) \n');
fprintf(fid,'Next z \n');

fprintf(fid,'swModel.Insert3DSplineCurve (False) \n'); % connect the points from 0

to 20 to gebarate a reference line

fprintf(fid,'Set swGuideCurve = swModel.FeatureByPositionReverse(O) \n'); %

define the reference line position in Feature object
Generate propeller blade by loft
function

fprintf(fid,"'create loft \n'); %create loft to generate a propeller blade

fprintf(fid,'swModel.ClearSelection \n'); %clear anything currently selected

116
 fprintf(fid,'"select profile curves \n'); % select all the foil sections first
fprintf(fid,"'mark = I \n');

fprintf(fid,'For z = 0 To 20 \n');% the order of selecting foi sections is from 0 to 20

fprintf(fid,'bRet = swFeatCurve(z).Select2(True, 1) \n');
fprintf(fid,'Next z \n');

fprintf(fid,"'select guide curves \n'); % select the reference line second

fprintf(fid,"'mark = 2 \n');
fprintf(fid,'bRet = swGuideCurve.Select2(True, 2) \n');

fprintf(fid,'swModel.InsertLoftRefSurface2 False, False, False, 1#, 0, 0 \n'); %loft to

generate a propeller blade
Hide the sketches for clean

fprintf(fid,'"hide profile curves \n') %hide the sketches for clean

fprintf(fid,'swModel.ClearSelection \n');
fprintf(fid,'bRet = swFeatCurve(O).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(l).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(2).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(3).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(4).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(5).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(6).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(7).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(8).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(9).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(l O).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(1l).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(12).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(13).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(l4).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(l5).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(16).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(1 7).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(l8).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(l9).Select2(True, 0) \n');
fprintf(fid,'bRet = swFeatCurve(20).Select2(True, 0) \n');
fprintf(fid,'bRet = swGuideCurve.Select2(True, 0) \n')
fprintf(fid,'swModel.BlankRefGeom \n')

fprintf(fid,"'hide the 3D sketches since they have a lot of points \n');

fprintf(fid,'"note that ModelDoc2::BlankSketch will only hide one sketch \n');

117
 fprintf(fid,'For z = 0 To 20 \n');
fprintf(fid,'swModel.ClearSelection \n');
fprintf(fid,'bRet = swSketch(z).Select2(True, 0) \n')
fprintf(fid,'swModei.BlankSketch \n');
fprintf(fid,'Next z \n');
fprintf(fid,'swModel.SetAddToDB False \n');
fprintf(fid,'End Sub \n');
fclose(fid);

end %(END IF Make_Solidworks_flag)

118
Appendix B: SolidWorks Macro for the AlJV Propeller Geometry

119
Sub main()
Dim swApp As SldWorks.SldWorks
Dim swModel As SldWorks.ModeiDoc2
Dim swSketchPt() As SldWorks.SketehPoint
Dim swGuidePt(21) As SldWorks.SketchPoint
Dim swSketch(21) As SldWorks.Feature
Declare all the
Dim swFeatCurve(21) As SldWorks.Feature
objects as described
Dim swGuideCurve As SldWorks.Feature
as Appendix A
Dim z As Long
Dim bRet As Boolean
Set swApp = CreateObject("SidWorks.Application")
Set swModel = swApp.ActiveDoc
swModei.SetAddToDB True
swModei .CiearSelection
'start 3D sketch
Start to create the points that
swModei.Inseti3Dsketch - - --.
ReDim swSketchPt(40) on foil section number 1
Set swSketchPt( 1 ) = swModei.CreatePoint2 (0.041867, -0.022025, 0.050618)
Set swSketchPt( 2 ) = swModei.CreatePoint2 (0.040091, -0.016425, 0.052702)
Set swSketchPt( 3 ) = swModei.CreatePoint2 (0.036700, -0.012675, 0.053 727)
Set swSketchPt( 4) = swModel.CreatePoint2 (0.03301 3, -0.009267, 0.054419)
Set swSketchPt( 5) = swModei.CreatePoint2 (0.029138, -0.006084, 0.054866)
Set swSketchPt( 6) = swModel.CreatePoint2 (0.025122, -0.003082, 0.055116)
Set swSketchPt( 7) = swModel.CreatePoint2 (0.020983 , -0.000247, 0.055202)
Set swSketchPt( 8) = swModei.CreatePoint2 (0.016733 , 0.002427, 0.055149)
Set swSketchPt( 9) = swModei.CreatePoint2 (0.012358, 0.004913 , 0.054983)
Set swSketchPt( 10) = swModei.CreatePoint2 (0.007883, 0.007241, 0.054725)
Set swSketchPt( 11 ) = swModel.CreatePoint2 (0.003254, 0.009329, 0.054408)
Set swSketchPt( 12) = swModei.CreatePoint2 (-0.001498, 0.011218, 0.054050)
Set swSketchPt( 13) = swModei.CreatePoint2 (-0.006356, 0.012935, 0.053665)
Set swSketchPt( 14) = swModei.CreatePoint2 (-0.011298, 0.014511, 0.053261)
Set swSketchPt( 15 ) = swModel.CreatePoint2 (-0.016313, 0.015962, 0.052844)
Set swSketchPt( 16) = swModei.CreatePoint2 (-0.021401, 0.017307, 0.052419)
Set swSketchPt( 17) = swModei.CreatePoint2 (-0.026529, 0.018548, 0.051993)
Set swSketchPt( 18 ) = swModei.CreatePoint2 (-0.031643, 0.019731, 0.051556)
Set swSketchPt( 19) = swModei.CreatePoint2 (-0.036749, 0.020891, 0.051097)
Set swSketchPt( 20) = swModei.CreatePoint2 (-0.041847, 0.022055, 0.050605)
Set swSketchPt( 21 ) = swModei. CreatePoint2 (-0.041888, 0.021994, 0.050631)
Set swSketchPt( 22) = swModei.CreatePoint2 (-0.037971, 0.019019, 0.051823)
Set swSketchPt( 23) = swModci.CreatePoint2 (-0.034044, 0.015989, 0.052836)
Set swSketchPt( 24) = swModei.CreatePoint2 (-0.030132, 0.012909, 0.053672)
Set swSketchPt( 25 ) = swModel.CreatePoint2 (-0.026269, 0.009786, 0.054328)
Set swSketchPt( 26) = swModei.CreatePoint2 (-0.022426, 0.006638, 0.054802)
Set swSketchPt( 27) = swModel.CreatePoint2 (-0.018545, 0.003514, 0.055090)
Set swSketchPt( 28) = swModel.CreatePoint2 (-0.014618, 0.000456, 0.055200)

120
Set swSketchPt( 29) = swModei.CreatePoint2 (-0.010628, -0.002503 , 0.055145)
Set swSketchPt( 30) = swModel.CreatePoint2 (-0.006555, -0.005324, 0.054945)
Set swSketchPt( 31 ) = swModel.CreatePoint2 (-0.002373, -0.007967, 0.054624)
Set swSketchPt( 32) = swModei.CreatePoint2 (0.001943, -0.010395 , 0.054215)
Set swSketchPt( 33) = swModel.CreatePoint2 (0.006341 , -0.012684, 0.053725)
Set swSkctchPt( 34) = swModci.CrcatePoint2 (0.01 0845, -0.014799, 0.053182)
Set swSketchPt( 35) = swModei.CreatcPoint2 (0.015441, -0.016765, 0.052595)
Set swSketchPt( 36) = swModei.CreatePoint2 (0.02013 7, -0.018573, 0.051984)
Set swSketchPt( 37) = swModei.CreatePoint2 (0.024951 , -0.020205, 0.0513 7 1)
Set swSkctchPt( 38) = swModcl.CreatePoint2 (0.029920, -0.021619, 0.050793)
Set swSketchPt( 39) = swModel.CreatePoint2 (0.035141 , -0.022703 , 0.050318)
Set swSketchPt( 40) = swModcl.CrcatcPoint2 (0.041867, -0.022025, 0.050618)
Set swGuidePt(O) = swSketchPt(1)
'exit 3D sketch
swModei.Inseti3 OS ketch
Connect all
Set swSkctch(O) = swModel.FeaturcByPositionReverse(O)
points to
swModei.ClearSelection
generate the
For z = 22 To 39
f()i) section
bRet = swSketchPt(z).Select2(True, 0)
and define
Next z
the point
For z = 1 To 20
position for
bRct = swSkctchPt(z).Select2(True, 0)
the reference
ext z
line
swModcl.Inscti3 DS plineCurve (True)
Set swFeatCurve(O) = swModel.FeatureByPositionReverse(O)
swModei.CiearSclection

This step is repeated until get 21 foil sections

'create guide curve Create reference line for

For z = 0 To 20 lofting. The detailed each step
bRet = swGuidcPt(z).Select2(True, 0) is described as Appendix A
ext z
swModci.Inscrt3DSplineCurve (False)
Set swGuideCurve - swModcl.FcaturcByPositionRcvcrse(O)

121
'create loft
swModel.ClearSelection
'select profile curves
'mark = 1
For z = 0 To 20
bRet = swFeatCurve(z).Select2(True, 1)
ext z
'select guide curves
'mark = 2
bRet = swGuideCurve.Select2(True, 2)
swModel.lnsertLoftRefSurface2 False, False, False, 1#, 0, 0
'hide profile curves
swModei.ClearSelection Hide all ofthe sketches for
bRct = swFcatCurvc(O).Sclcct2(Truc, 0) clean
bRet - swFcatCurvc(1 ).Select2(True, 0) L . , __ _ _ _ _ _ _ _ __ __,

bRct - swfeatCurve(2).Selcct2(True, 0)
bRct = swl·catCurvc(3).Sclect2(True, 0)
bRct = swFcatCurve(4).Select2(True, 0)
bRct = swFcatCurvc(5).Select2(True, 0)
bRet = swFeatCurve(6).Select2(True, 0)
bRct = swFeatCurve(7).Select2(True, 0)
bRet = swFeatCurve(8).Select2(True, 0)
bRet = swFeatCurve(9).Select2(True, 0)
bRet = swFeatCurve(l O).Select2(True, 0)
bRet = swFeatCurve(1l).Select2(True, 0)
bRet = swFeatCurve(12).Select2(True, 0)
bRet = swFeatCurve(13).Select2(True, 0)
bRet = swFeatCurve(14).Select2(True, 0)
bRet = swFeatCurve(15).Select2(True, 0)
bRet = swFeatCurve(16).Select2(True, 0)
bRet = swFeatCurve(17).Select2(True, 0)
bRet = swFeatCurve(18).Select2(True, 0)
bRet = swFeatCurve(l9).Select2(True, 0)
bRct = swFcatCurve(20).Select2(True, 0)
bRet = swGuideCurvc.Select2(True, 0)
swModel.BlankRefGcom
'hide the 3D sketches since they have a lot of points
'note that Mode1Doc2: :BiankSketch will only hide one sketch
For z = 0 To 20
swModei.CiearSelection
bRet = swSketch(z).Select2(True, 0)
sw Model. BlankS ketch
ext z
swModei.Seti\ddToDB False
End Sub

122